Train Your Dog: Reinforcement and Punishment in Operant Conditioning
Primary Reinforcement Basics: Food
Food is a primary reinforcer, meaning it relates directly to the dog’s biology. Food is what dogs need to survive. For thousands and thousands of years, dogs have related food (and the ability to eat), as a reward for their successful behaviors while hunting.
Food is also the easiest for us humans to manipulate in our dog’s environment, as we tend to have total control over feeding. While we may dole out praise and toys when playtime comes around, generally—and what we, in fact, do NOT want to do —is hand out food tidbits for free. Even mealtimes can become valuable training exercises for our dogs.
Reinforcement can also be in the form of play, praise, and attention. These are some of the typical things we think about when we think of rewarding our dogs as they relate to us.
Reinforcement can also be life rewards, for example, walking through a door or a gate to an exciting new environment, getting to sniff that particularly interesting tree during a walk, getting to play with another dog at the dog park, etc..
Keep in mind, there are some factors you’ll want to consider about the reward when deciding how to reinforce your dog’s behaviors, such as the reward’s consequent level of excitement and its value. As a rule of thumb: food rewards are the best rewards.
Positive vs. Negative Reinforcement
To refresh, positive reinforcement in this context means to add, and negative in this context means to remove.
So here is an example of a "negative" used in dog training: you are playing with your dog and a play toy and the game is getting too rough. You end the game and remove the toy as a message to the dog that this behavior ends the fun. This is an example of negative punishment. We are removing something and the result is punishing.
Issues With Positive Punishment
The overarching number one issue with positive punishment is neutral. Simply put, positive punishment has a lot of rules.
1) It Must Be Punishing to the Dog
What is NOT inherently always punishment is what we as humans think is punishing. Take the example of a dog who jumps up. Well-meaning owners attempting to decrease this behavior will push the dog away with a stern voice, indicating in our best human language that we don’t appreciate the behavior and the dog should keep its distance. Dog’s don’t speak human languages. In dog language, the act of pushing the dog away could be a sign to the dog that you are actually now engaging in play.
Many dogs of all breeds and some breeds more than others engage in a lot of body contact during play. The act of pushing and shoving and being pushed and shoved is all part of the game. So for the dog who continually jumps up on family and strangers alike and is met with a pair of playful hands, not only is this action not punishing to the dog, it’s reinforcing! We go back to our number one rule: punishment must be punishing to the dog.
2) Punishment Must Work to Decrease the Behavior
Punishment means something that works to decrease the likelihood of a behavior in the future.
So what if it doesn’t work? What if, despite our best efforts, the dog’s undesired behavior continues? What when, after our loudest vocal reprimands, the dog is still barking? What when, after years of popping our dogs on choke chains, the dog continues to pull on leash?
Many dog trainers and operant conditioning learning theorists alike may go on to describe the repetition of unsuccessful "punishments" to our dogs as unnecessary—after all, if it is not working to decrease the likelihood of the undesired behavior, why keep at it? Sometimes, punishments may even be labeled as abuse. Changes from an effective punishment should be seen almost immediately and within only the first few repetitions should a noticeable decrease in the likelihood of a behavior be observed. If it’s not working, it’s not proper punishment.
3) Punishment Must Not Be Associated With Us
We want training to be fun for our dogs and ourselves. We want a dog who’s engaged and willing to work for us. We don’t want our dogs to learn to fear us by having them associate us with the punishment they may receive.
Dogs, not unlike other intelligent species, are learning machines. Dogs are experts in the field of making associations with things good and bad, and learning by consequence of their behaviors which ones to repeat again to their benefit in the future.
Oftentimes, we as humans do not think about what associations could be being made in our dog’s mind. In fact, we barely have a handle on associations our own brains make.
Example: Your new puppy has just piddled on the living room floor. A classic scenario is then that the owner reprimands the puppy, perhaps by vocal reprimand, putting their hands on the dog in some form, rubbing their nose in the mess, or angrily directing them outdoors. What associations could the puppy have made in this scenario? Surely we have made it clear to the puppy that making a mess on the living room floor is not the thing to do (that they should go outside instead). In fact, in the eyes of the puppy, all that they may have learned is that going to the bathroom in front of you is dangerous! So the next time they go to the bathroom they do it in a separate room, in a corner, behind a door, etc.. If you really want to be thorough with this method of training, despite its associative drawbacks, you would have to reprimand your puppy for peeing inside the house not just everywhere, but every time at the time of it happening.
There are better ways of changing your dog’s undesirable behavior that does not involve our dogs associating us with danger or adding anything otherwise unpleasant into your dog’s life.