Knowing how to use reinforcement and punishment is, without a doubt, a very important part of working and living with animals. All good trainers and behaviourists use both reinforcement and punishment to condition animals, the key being the kind of reinforcement and punishment used.
In ethology (the study of behaviour) and psychology:
– Reinforcement refers to anything that increases the likelihood of a behaviour repeating itself
– Punishment refers to anything that decreases the likelihood of a behaviour repeating itself.
Both reinforcement and punishment can be either positive or negative. Unlike in everyday English, in ethology, ‘positive’ and ‘negative’ do not refer to ‘good’ and ‘bad’; they instead take on a more arithmetical definition. Positive refers to adding something, while negative refers to taking it away.
– Positive Reinforcement: adding something pleasant (said appetitive stimulus) in order to increase the likelihood of a behaviour repeating itself, for instance, giving a treat to your dog when he sits will increase the likelihood of him sitting again.
– Negative Reinforcement: taking something generally unpleasant (said aversive stimulus) away, a prime example is the fairly out-dated practice of ‘ear-pinching’ to force-teach a dog to hold an object in his mouth: The trainer pinches or twists the dog’s ears until he accepts to hold the training bumper in his mouth.
– Positive Punishment: adding an aversive stimulus in order to decrease the likelihood of a behaviour repeating itself (“correcting” a dog using an e-collar/electric collar if he goes to chase after a cat).
– Negative Punishment: taking an appetitive stimulus away in order to decrease the likelihood of the behaviour repeating itself. This can be seen when you leave the room and shut the door when your dog exhibits attention-seeking behaviour, like whining or pawing at you. Her pushiness is punished by the appetitive stimulus (you) being taken away.
Needless to say, the sooner the reinforcement or punishment is given, the more likely it is to be effective. Too late and you’re reinforcing or punishing something completely different. This is one reason clicker training is so successful: it gives the trainer the possibility to praise a behaviour the instant it happens. Click here for more on clicker training.
The value of praise and punishment depends on what the dog likes and dislikes. My Border Collie mix will fulfill a request more readily if I’m holding a ball rather than a treat. Some dogs won’t react at all to a ball. Punishment is the same: your dog jumps all over you and your reaction is to tell her, “No!” in an angry tone and push her down, some dogs will cower in response, others will only get more excited and possibly jump up even more! (It should be noted that this seemingly visceral reaction to dogs jumping up is ineffective and should be avoided.) Humans also have this innate reflex to pat a dog on the head as a show of affection. I’ve yet to see a dog (or any animal, for that matter) appreciate being repeatedly smacked on the head, although most seem to patiently tolerate it. Since the response to such a “reward” is typically subtle (lip licking and turning the head away) it will often elude the inexperienced handler. This is a prime example of why it is important to correctly interpret what an individual deems to be of value.
The environment you find yourselves in can greatly alter what your dog considers a reward: my German Shepherd mix will stare longingly at me for a tasty treat if we’re training at home, but will adamantly refuse one on our walks, where the chance to sniff the ground is the greatest positive reinforcement he could possibly ask for!
The interval of administration is also important. After a behaviour has been ‘captured’ (put on cue), the trainer can then increase the times when the dog doesn’t get a primary reinforcement and instead only a secondary one; very much like a slot machine: your dog will always be hopeful that this time they get a treat. Giving a treat for every single little thing for too long is detrimental to the training process. Think of the difference between a vending machine and a slot machine: how many more dollars are you going to put in a vending machine if the first time you do it, nothing comes out? Now how about a slot machine? Also, the more irregular the primary reinforcements are, the better. Casinos thrive on the fact that gamblers will always be hoping that the next spin, hand or draw is their lucky one! Remember, we’re animals, too!
It should be noted that the slot machine principle does not apply to punishment, only praise. Sporadically and inconsistently correcting an undesired behaviour will have the opposite effect on your training: your dog will wonder if this time he barks at you for attention, you’ll ignore him or not, and thus be more likely to take his chances and do so. Dogs are phenomenal risk takers, they adamantly adhere to the philosophy that you’ll miss 100% of the shots you don’t take!
It is important to make a distinction between primary and secondary reinforcement/punishment. A primary reinforcement/punishment is one that is inherently appetitive/aversive (respectively) to the animal. The value of said stimulus doesn’t need to be taught. Again, what is perceived as a reward or reprimand varies from individual to individual, but generally treats and toys are primary reinforcements and pain is a primary punishment. If you have a well-loved dog, one who is often cuddled and doted upon, he will not consider a scratch behind the ear ‘good enough’ to be primary reinforcement. Dogs who live in loneliness will more likely consider any affection sufficient primary reinforcement.
A secondary reinforcement or punishment has been given along with a primary one for some time, and later, when given alone, is symbolic of said primary reinforcement/punishment; a rain check, so to speak. Typical secondary reinforcements are the sound of a clicker or verbal praise; a typical secondary punishment is a verbal reprimand.
Last but not least, life rewards are also a boon to your training. Life rewards are all things your dog looks forward to in a specific context. Example: wanting to go outside when he needs to pee. Ask for something (ex.: a sit, down or stay) before you pet your dog when he wants attention, before you feed him dinner or let him outside. Cuddling, dinner and getting to go outside are all rewards for your dog, use them!
Take the time to really observe how your dog reacts when you reward her. Watch to see if she looks excited to continue training, or a little unenthused. Knowing what your dog would jump through fire for has the potential to radically improve the outcome of your training.
Good luck, and happy training!