How Dogs Learn by Consequences: Operant Conditioning in Dog Training
An in-depth look at how dog training works
Photo by Crew
When we train our dogs sometimes we question our efficacy and some other times the efficacy of whatever method we follow. While there are methods that suck, most of the times the fault of an unresponsive dog is ours.
That’s when an in-depth understanding of the psychology behind dog training becomes necessary.
Operant conditioning: the definition
Operant Conditioning is a learning model that describes how intelligent beings change their behavior according to the consequences that follow.
For example, a student (a human student) is more likely to take part in a class that offers incentives. If good grades are also followed by rewards, the student is set up for success. Adding a pleasant outcome to a voluntary response motivates the student to do the work. She is now more likely to find success.
In the same fashion —back to the context of dog training— we introduce a consequence when a behavior occurs, in a way that makes sense to the dog.
There are two ways of doing that. One way is to focus on the behavior we want to increase (reinforcement). The other way is to focus on the behavior we want to decrease (punishment).
We decrease or increase a behavior by adding (positive) or removing (negative) a stimulus. The addition or removal of a stimulus is the consequence that dogs learn to associate with their actions, and this is how learning occurs.
It will make more sense if you keep on reading.
The four types of operant conditioning
- Positive reinforcement. A pleasant stimulus follows the desired behavior. E.g., we give praises to the puppy that just pooped in the designated area.
- Negative reinforcement. An unpleasant stimulus is removed when the desired behavior occurs. E.g., the handler restrains the dog until the dog is calm, when the restraint is removed. Negative refers to something that we remove (the restraint) when the desired behavior occurs (the dog is calm).
- Positive punishment. An unpleasant stimulus follows the undesired behavior. E.g., the dog jumps on visitors and the handler shouts to the dog. Again, positive refers to the addition of shouting.
- Negative punishment. A pleasant stimulus is removed when the undesired behavior occurs. E.g., the dog eats poop from the ground and the playtime is over. Negative refers to the playtime being over.
*In operant terminology, the words positive and negative are used in their mathematical sense. They mean addition and abstraction. Punishment, in the context of operant conditioning, is a technical term that means the introduction of a consequence after a particular behavior occurs to make it less likely to reoccur. While this leaves room for abuse, as it has, it doesn’t mean that punishment, if wisely used, is abusive.
The four types in-depth:
This is the most dog-friendly way to train your dog.
Positive reinforcement is a powerful method of teaching that introduces pleasant stimuli after a desired behavior occurs. In dog training, it works by rewarding a dog for offering a behavior that we want to encourage. Rewards may include food, praises, toys, games, etc.
Positive reinforcement relies on the simple notion that dogs are more likely to repeat behaviors that they associate with pleasant consequences. Basic cause-and-effect stuff.
The word positive, as in math, refers to something being added (the reward), and reinforcement to the behavior we want to encourage. For example, with a soft and firm voice say ”come”, and give a treat when the dog comes. The treat is the added stimulus and the successful response to our command ”come” is the behavior you want to reinforce.
See the sequence of events? Cue, response, and reward.
Positive reinforcement values kindness and cooperation. Desirable behaviors followed by desirable consequences. Quite positive, indeed.
Negative reinforcement happens when we remove an aversive the moment the desired behavior occurs. In other words, dogs learn to behave in ways that release them from discomfort and even pain.
Take prong collars, for example. Prong collars give pressure to the neck of a dog and release only when the dog stops pulling. This makes dogs stop the pulling. Dogs learn that walking next to you releases the discomfort. Then, they learn to actively avoid it.
There are two ways that negative reinforcement works. One is when the trainer removes the aversive stimulus (escape). Second is when the dog actively avoids a known unpleasant outcome (avoidance.)
Examples of negative reinforcement:
- The trainer shocks the dog with an e-collar when she walks away and then commands her to ‘come’. The trainer releases when the dog comes. The dog learns that coming after the cue/command ‘come’ is going to release her from pain (escape.)
- The dog stays around his trainer because walking away causes discomfort (avoidance).
- The dog knows that barking to strangers makes his people mad, so he prefers to avoid it (avoidance.)
Training with aversives requires you are very cautious and experienced. I never purposely give pain to my dog, and I don’t recommend it to you either. You risk ruining your relationship with your dog.
In any case, if you are convinced that this is the way-to-go, do not use negative reinforcement for long, and don’t make the aversive stimuli look like it comes from you. Make them look like natural consequences to your dog’s actions. Best thing, get professional help.
Now, to avoid the common confusion that makes people mistake negative reinforcement with punishment, keep reading. Keep in mind that with negative reinforcement we aim to increase a behavior, while with punishment we aim to decrease it.
A more familiar concept, punishment, happens when we introduce an unpleasant consequence when an unwanted behavior occurs. Positive punishment is when we add an unpleasant stimulus when a dog does something we do not approve or doesn’t do what is being asked.
For example, the dog walks ahead and the collar tightens. The pressure is the aversive stimulus that punishes the dog for walking ahead. That’s positive punishment. With time, the dog will learn to avoid this stimulus but not walking ahead. And that’s negative reinforcement (avoidance) in action.
Other examples of positive punishment are e-collars, shouting, prong collars, beating, showing anger, restrictions, and more.
It’s easier for most people to grasp the concept of positive punishment compared to the less familiar negative reinforcement and negative punishment. It has been the way to teach and manage children for centuries. When people started training dogs methodically, about a century ago, they quickly picked up what behaviorists knew at the time.
Training methods that stem from or use negative reinforcement and positive punishment is the so-called dominance training, and the alpha dog training. The latter is somehow still a thing.
Negative punishment happens when we remove a stimulus that the dog values as a result of his/her unwanted behavior.
For example, when my dog eats poop at the park, I express verbal disapproval, and immediately she is restricted for a minute to a designated penalty area, or even sometimes our walk is over. Her freedom or playtime is over (removed) the moment she engages in a highly disapproved behavior.
Negative punishment is the dog-friendly way to communicate your disapproval and correct your dog. Dogs learn that unpleasant consequences are part of life.
Key-points and clarifications:
1. The difference between operant and classical conditioning
Classical conditioning is learning by association. Dogs observe the correlation between two or more events and learn to act accordingly. For example, my dog knows that the sound of the alarm in the morning (first event) means that we are getting up (second event), and a new day of fun begins. She knows that when I make breakfast she gets half an apple, that getting her leash means we are going out, that coming back means she gets to get cleaned, and she knows very well which sounds mean meal-time.
The main difference between the Operant and Classical conditioning is this: In OC we, the handlers, link behaviors to events (the introduction of stimuli), while in CC two events are naturally paired to each other. In OC the dog voluntarily chooses to offer a behavior in order to enjoy a pleasant consequence or give up a behavior to avoid an unpleasant one, while in CC two independent events are linked and cause the dog to act accordingly.
2. Reinforcement focuses on the desired behavior, while punishment focuses on the undesired
Another way to categorize the four types of Operant Conditioning is by their intent. With positive and negative reinforcement we want to teach new behaviors and with punishment we want to stop an established one.
Punishment doesn’t guide towards what’s good. Punishment stops what’s ‘bad’. While this is important (when not complying with known commands gets dangerous), guiding towards what’s desirable should come first.
3. Behaviors treated with punishment are oppressed, not resolved
Punishing alone won’t resolve much. It is a momentary correction that with time and repetition forms fear-related compliance. Punishment relies on fear. Dogs will only comply when the punisher is there. The behavior is not forgotten unless they learn otherwise.
For example, merely punishing a dog for aggression towards strangers doesn’t resolve the issue, it oppresses it. Creating anew a more positive experience for the dog around people will resolve it.
4. Reinforced behaviors tend to extinct if not continuously reinforced
Even established behaviors extinct if not continuously reinforced. When a behavior is somewhat established, move on to the next thing you want to teach your dog. For the known behaviors(commands if you prefer), employ real-life rewards.
Real-life rewards are a convenient way to reinforce behaviors out of the context of formal training. Everything that your dog likes during a normal day can be a real-life reward for a behavior you want to reinforce.
When my dog Aria learned to sit, we moved on to the next command/signal. Training devices are reserved for new goals. ”Sit” was well established. ”Sit”, and other known commands, now happen before we open the balcony door for her before we put food in her bowl, before she gets on the couch, etc. Everything she loves is a chance to reinforce her responses.
5. Positive reinforcement and negative punishment go well together
Training with positive reinforcement and correcting with negative punishment minimizes the chances that out-of-control situations will ever come.
This combination ensures you leave cruelty out. With positive reinforcement, we build relationships based on cooperation and kindness. With negative reinforcement, we show that unpleasant consequences are part of everybody’s life. The latter is a law of nature. We don’t argue with Nature.
6. Positive reinforcement should be your primary method of training
If you want to be your own dog’s trainer, which is very doable, positive reinforcement is the way-to-go. Positive Reinforcement is the method that most dog trainers in the world recommend as everybody’s primary method of training. It is the only method that guarantees a smooth symbiosis with our dogs.
Positive reinforcement regards and respects a dog’s enthusiastic and curious nature. It acknowledges a dog’s talent to find fun in everything around them. With positive reinforcement, you won’t have just an obedient dog, but a partner in fun.
Dogs learn to offer the desired behavior because they want to and not because they fear bad consequences.
Positive reinforcement is, no doubt, the best option for strengthening the bond we share with our dog(s). There is undeniable evidence for that. It’s our way to speak the Dog language and dogs’ only chance to be their best versions.
Positive reinforcement should be our first choice of training.
Remember that the four quadrants are not cut and dry. Often it is difficult to say if restricting your dog’s freedom is positive or negative punishment. Your best bet here is to make sure your dog gets the physical and mental daily exercise she/he needs. Honestly, there is no better way to ensure a smooth symbiosis with your dog.
- Dog training methods: Their use, effectiveness and interaction with behaviour and welfare.
- Purely Positive, Force Free, and Science.
- The Evolution of Modern-Day Dog Training & Obedience