Let's start by understanding that a dog's brain works much the same way as does our brain, a dolphin, a bear or even a chicken. We all learn through a process of experimentation whereby we offer up behaviors and depending on the consequence that behavior brings, we decide to repeat the behavior or not. Those behaviors that bring a pleasant consequence are those that we will likely repeat, whereas those that bring an unpleasant consequence are those behaviors that we will offer less often or certainly not at all.
This process includes 4 basic laws of learning where behaviors create consequences. These consequences either come directly from the environment or perhaps from another dog or a human. Trainers understand and utilize these principals in order to manipulate a dog's brain. We then can teach the dog to perform a task and to maintain harmony in a social environment. Actually, as this process continues, the laws blend in a synergistic way with one another. Highly skilled trainers can be seen working very smoothly and effortlessly with dogs as they shape behaviors.
When you control the consequences, you control the behavior!
Behavior is driven by the consequence it brings!
The Laws of Learning
Positive Reinforcement
In behavioral terms, a positive reinforcement is when something pleasant is added as a response to an animal's environment that increases the likelihood that the dog's behavior will repeat and increase in frequency. Science has demonstrated that behavior that is followed by a pleasurable response is most often repeated. Therefore when your dog sits or downs and receives a reward for that behavior, he is apt to do it again and again. When dogs bark and get attention, they bark again. When dogs jump on counters and find food, they jump again. When dogs jump on people and get attention, they jump again.
Negative Punishment
In behavioral terms, a negative punishment is when something pleasant is withheld while undesirable behavior exits. Basically the handler deprives the dog of rewards like food, fun, praise, attention or an activity while undesirable behavior presents itself. For example, as long as the dog barks, jumps or pulls on the lead, the handler will not give the dog what he wants, until he is quiet and gentle.
Positive Punishment
In behavioral terms, a positive punishment is when something unpleasant is added to an animal's environment that suppresses that animal's behavior. Since this type of punishment often results in discomfort, pain or fear, dogs will begin to avoid it at all cost. This avoidance behavior is considered the fallout of positive punishment. This avoidance is most easily witnessed when dogs are introduced to underground electric fencing, but in this case it is the desired effect. However, when the handler responds with a harsh correction whenever the dog is excited around other dogs on the street, perhaps the fallout might be that the dog begins to dread seeing other dogs on the street for fear of a harsh correction and the excitable and anxious behavior worsens.
Negative Reinforcement
In behavioral terms, a negative reinforcement is when something unpleasant continues as long as undesirable behavior presents itself. Basically nagging at a dog with corrections until the dog shuts down and can be redirected towards a more desirable behavior. A handler might decide to continue to give electric stimulation, or leash corrections or shout "NO" over and over again until the dog stops moving forward, or misbehaves. The dog is supposed to learn that he can prevent unpleasant responses like these by choosing an alternate behavior until there is an absence of discomfort.
Analogies for the support of reward based training
# 1
As drivers, we all know that if we go faster than the speed limit sign dictates, we will be punished. This punishment is applied by a police officer. When we exceed the speed limit, we are pulled over, given a ticket, we get points added to our license, our insurance goes up and we must pay a fine. Sounds horrible, right? Then why do people still speed? Why do we have repeat offenders? Consider this...
- We can speed when no one is looking, so the behavior continues.
- We can speed when there is an emergency, so the behavior continues.
- We can ask the judge to lower the fee when we have financial problems, so the behavior continues.
- We can negotiate with the officer at the time, so the behavior continues.
- We can take defensive driving courses to lower points on our license, so the behavior continues.
If we want to PUNISH effectively, we have to be sure to apply that "something" timed specifically to the exact moment of the speeding behavior. Both reinforcement and punishment need to be timed specifically to the behavior otherwise they will not influence that behavior. In this case, there needs to be an exact association with the punishment being the result of the behavior. Keep in mind that the punishment must also be impactive enough so that the behavior would never be repeated! Consider how, instead of being ticketed for speeding, that instead officers could pull speeders aside, ask them to get out of their vehicle, and BLOW UP THEIR CARS! You wouldn't have a car anymore! That would surely be a deterrent.
The problem with this form of punishment is the FALLOUT... we already get nervous when police are driving near us on the road even while we are behaving appropriately. We often slow down and try to avoid the police car. Imagine how we'd feel around policemen if we thought there was a chance of them blowing up our CARS!
Using the science of behavior and the power of positive reinforcement, consider this alternative. As we are driving around and NOT speeding, we would be pulled over by the police and given 50 bucks! Every time we are NOT SPEEDING we get another 50 bucks! Wouldn't we all slow down on the roads? We'd begin to "offer" behaviors, like driving extremely slowly, or traveling exactly at the speed limit. We'd experiment to see exactly which behaviors resulted in the most rewards! Wouldn't we all choose to get up early or stay out later, exhibiting non-speeding behaviors, just to get more money? That is of course only if money is what we needed at the time! Suppose you already were very wealthy, and didn't need 50 bucks. Consider the kind of reward that would influence those people in order to manipulate the behavior? That is the exciting and rewarding part of reward-based training. You must learn how your dog thinks and feels about various situations and rewards and how to inspire him. You need to offer various types of rewards for different dogs and different situations.
#2 Imagine that you've arrived in a foreign country and couldn't speak the language.
You get off the train onto a platform where a man with a red hat begins to tug on your sleeve quite harshly yelling "ZUF.. ZUF.. ZUF!!! You begin to move along and eventually get to the other side of the platform. You are quite shaken up, and you have no idea why that happened, but you got through it.
The next day you arrive at the same station and you are a bit nervous as soon as you see the guy with the red hat. Again, you step off the train and he begins his harsh sleeve tugging shouting "ZUF.. ZUF.. ZUF" !!! This time you brace yourself because you know it will be over soon and you get to the end of the platform.
Day after day this happens and you wish you could get away from the guy with the red hat. You realize that you can't so you continue to endure it and actually become numb to the whole ordeal.
Now imagine the same train and the same platform. This time you get off onto the platform and a guy with a red hat greets you with a cheery voice and hands you some candy. He continues to walk along side you adding candy with every other step along with a lovely happy voice and accompanies you across the platform to the other side.
The next day this scenario is repeated day after day. You have no idea why but you look forward to seeing the guy with the red hat.. you actually look for him in the crowd.
Then one day the guy with the red hat begins to withhold candy and his cheery voice is quiet. He begins to only offer candy and his happy voice when you walk on a certain side of the platform, but not when you are on another side.
Another day the guy with the red hat starts to offer candy and a cheery voice only when you walk very slowly and only when you are on a certain side of the platform.
Now I've given you something to think about!
Here is an important aspect of reward based training
Imagine that you are playing poker for the first time.
- You lose your first hand, would you continue to play?
- You lose the second hand, would you continue to play?
- You lose the third hand, would you continue to play?
- You lose the fourth hand, would you continue to play?
- You lose the fifth hand, would you continue to play?
Most probably NOT!
But we all understand why.
Suppose instead:
- You win the first hand, would you continue to play?
- You win the second hand, would you continue to play?
- You win the third hand, would you continue to play?
- You win the fourth hand, would you continue to play?
- You win the fifth hand, would you continue to play?
Most probably NOT!
Isn't this interesting? At some point you would not want to keep playing because you win all the time! You would have enough money and you would become quite bored because there is no challenge to winning. So, to keep you in the game, you certainly need to win at first to believe you can win, then lose periodically in a random win/lose fashion to keep it exciting. You then would begin to learn what you need to do in order to win.
That is what you want to create in your dog's mind. You want him to believe that he can win food, attention, praise and activities by executing certain behaviors. You want him to choose his behaviors and do so with confidence that certain behaviors do pay off!
Be careful NOT to live in the land of "NO" but rather, try to live in the land of "YES". You will see desirable behaviors built over time and undesirable behaviors will gradually diminish. You want to have a relationship with your dog built in trust, so that he is confident and eager to please. Be ready to reward his appropriate behaviors and you will see an eager, confident response to cues and an increase in reliability of behaviors that are built over time.
| 
