Treats, scolding, reinforcement, punishment

What works, how does it work, what does it mean?

Good Work Dog uses and promotes positive reinforcement, and discourages any training that relies on punishment. Most of you will understand that to mean we give dogs lots of treats, and we don’t hit the dog. But there are many more layers to it. This article will try to clarify as much as possible and explain why we take this position.

In the early 1900s, several scientists such as Pavlov and Skinner tried to find reliable, testable patterns in the behavior of humans and other animals. Why do we act one way or another, and how can we change someone’s behavior? Everyone has heard of “Pavlov’s bell”, where he would ring a bell and his test dogs would start salivating because they expect food. These scientists discovered that there really are specific ways to understand how we learn new behaviors; Skinner described it as four quadrants, and we still use that model today.

In this system, “positive” means we add or give something, and “negative” means we take something away. They don’t mean good versus bad. “Reinforcement” means it makes a behavior increase or repeat, and “punishment” means it stops or decreases a behavior.

  • Positive reinforcement (R+) means we gave the dog something they like enough to repeat whatever behavior got them the reward, hoping for more.
  • Positive punishment (P+) means we did something that hurts or scares the dog enough to make them stop a behavior, to avoid getting punished.
  • Negative punishment (P-) and negative reinforcement (R-) mean we took away something the dog likes or dislikes, to change the choices they make.

An example of negative punishment is taking away your kid’s video game because they did something you don’t like. Notice that even though it’s “punishment”, it’s not as severe as if you hit them or yelled at them. You’re thinking that if you take away the game, they will learn their lesson; but let’s be honest, it doesn’t really work very well. Hitting or yelling (P+) obviously has a bigger impact, it produces strong responses, very quickly. So for someone who just wants results right now, punishment seems like it works. But it also causes damaging side effects: it makes the dog or child distrustful, anxious, emotionally frozen, and more likely to have an outburst of reactive behavior later on.

Positive reinforcement pays the dog for their work, encourages them to keep working with us, and informs them exactly which behaviors get rewarded. Evidence shows that real, long-lasting learning occurs best with R+. The “negative” quadrants do work, but their results are weaker and the learning doesn’t stick long term. Positive punishment is effective but also very harmful.

Some trainers and organizations use a model called LIMA: Least Intrusive, Minimally Aversive. It describes the progression of steps a trainer will take, starting always with R+ and only moving to R-, P-, and finally P+ if they absolutely can’t get results at the previous stage. A criticism of LIMA is that if the trainer does a more careful job of setting the dog up for success, they should not have to use less effective or more harmful methods.

What does all this mean for the dog owner at home?

Obviously you’re not thinking about quadrants when your dog barks at the mail carrier and you say “hush! No barking!” Any time we shout, scold, or tsk-tsk at our dogs, it’s really just a way to express our own frustration. It doesn’t have a meaningful impact on the dog’s learning, even if they stop what they’re doing in the moment. It’s OK, get it off your chest, you are not really hurting anything–but it doesn’t help much either. They will bark again next time.

And what about when you tried giving the dog treats, and their bad behaviors didn’t stop? Isn’t that proof that R+ didn’t work? In reality it shows that the dog wasn’t set up to succeed. Maybe there were too many distractions, maybe the dog didn’t understand the game, or there was something in the setup that the dog found too upsetting; there are lots of factors to manage. Dog owners sometimes put their dogs in a situation where they will fail, and give treats in a way that doesn’t have a clear meaning to the dog, and then get frustrated that the treats didn’t work. Reach out to us so we can help you and your dog work more effectively.

What about leashes, fences, muzzles?

They aren’t inherently reinforcing or punishing unless we make them that way. Anything we use to limit a dog’s choices is called “management”. A closed door is management, and putting your hamburger up out of the dog’s reach is management. These stop the dog from certain behaviors, but they do not teach the dog how to make different choices. As soon as you leave the hamburger on a low table, the dog will eat it. All training plans start with management, such as having the dog on leash or in a safe space, but by itself that doesn’t cause learning. The quadrants only apply when behavior has changed as a result. But well-planned management is a big key to setting a dog up for success in training.

Consent and Feelings

Another huge element not clearly addressed by the quadrants is consent. Even training that looks like R+, where you ask for a behavior and give them treats, may turn into punishment if the subject cannot consent or opt out. Good Work Dog uses training games that actually ask the dog for consent. The games offer opportunities, repeatedly throughout any session, for the dog to choose whether to proceed. Working in an hour-long scheduled block of time can mean we have to take many breaks and restarts, and we use management to limit how far the dog can go with opting out, but always we ask and “listen” for an answer by observing the dog’s behavior choices and body language. When the dog says no, we change the activity or the environment to find where the dog may be more willing to opt in.

Additionally, how does the subject feel about the training work, both physically and emotionally? A dog may deliver a behavior for a treat, but they may be bothered or upset enough that the whole process becomes aversive. Maybe they have a sore joint, a bad tooth, or a skin infection; any physical discomfort, especially if they associate it with trauma or a long-standing condition, can change how they feel about doing the work. It can even “poison” how they feel about food or toys! The same goes for emotions: a dog may go along with your R+ style training in a performative way, but if they are actually feeling stressed or worried, the training can do more harm than good. You will not see the behavioral results you wanted, and the dog may even “act out” in new unwanted ways! We always have to take feelings into consideration, for any of this to work.

Why is R+ “the new way” and punishment the old way?

Quote from E. H. Richardson: The more one has to do with the canine race the clearer it is seen that as regards psychology there is little difference between it and the human race, and I much regret to say that it was my experience to find occasionally the canine "conscientious objector" among the recruits. There was, however, a convenient method of dealing with the offenders which unfortunately is not available for human beings-an excellent lethal chamber at Battersea!
Lt. Colonel E.H. Richardson, “British War Dogs, Their Training and Psychology” 1920

Historically, dog training was mainly a practice of farmers, fishermen, and others who used dogs as working helpers. They relied heavily on breeding their dogs to instinctively perform necessary tasks, and taught them by working side by side in their trades. They bred smaller, gentler dogs for indoor companions, and larger, more aggressive dogs for guarding property. They did not take dogs bred for one job and try to train them for anything else.

Leading up to WWI, the German and French militaries took working breeds and trained them for wartime jobs, imposing the same strict discipline as boot camp for soldiers. Britain then came up with a war dog program to compete with the Germans, and military dogs such as Rin Tin Tin became hugely popular in movies and stories, so military-style training became the worldwide standard. In America, veterans returning from WWII bought a home with their government benefits, settled down to raise a family, and got pet dogs. Suddenly large numbers of dogs were expected to live in suburbs and cities with a huge new generation of children, the “baby boom”, which led to increasing expectation that the dogs would behave appropriately in this new lifestyle. A generation raised on wartime culture demanded obedience. This style of training remained the established norm through the middle of last century, despite the earlier discoveries of behavior science.

As later generations grew up without so much military influence, they started to ask whether there was a kinder, more encouraging way to raise both children and dogs. The evidence had already been laid out by Skinner and Pavlov, showing that reinforcement was more effective and less harmful, so more and more dog trainers and owners have transitioned to this approach. It’s still a big public debate, because people who were themselves trained with a violent belief system are often unable to accept that what they were taught might not be the best choice. Change is hard! But we believe in encouraging and rewarding you for making better choices as you learn how, even if it takes some time. Good work, dog!

Citations: