In the terminology of learning theory, positive does not always mean good. Actually it never means good. Likewise, negative does not mean bad. Also, reinforcement is not always about giving the dog something she wants. And punishment is not always about hurting, intimidating or confining her.
Got that straight?
If your head already hurts, skip ahead and watch the movie. That’s why I made it. The purpose of this post is to introduce a movie that illustrates the four processes of operant learning with multiple examples, in hopes of clarifying the bigger concepts with those examples. (Don’t worry; Feisty the stuffed dog stands in when bigtime aversives are used.)
Anybody still reading? Good.
To continue: Most of us have heard the term “positive reinforcement” and have a notion of what it means. The dog sits, you give her a cookie. That is by no means the whole picture, but it’s a start. But when the other terms start marching out, things get dicey. Negative reinforcement, negative punishment, and what, positive punishment? Isn’t that an oxymoron?
There is a certain amount of confusion that we just have to live with on this subject because of the terminology that B. F. Skinner chose. In learning theory, positive means something is added. Negative means something is removed. They do not mean good and bad, happy and sad, or moral and immoral. Susan Friedman points out that in the context of mathematics, no one mistakes a plus sign to mean something happy or morally desirable. It just means you add that thing. Likewise with a minus sign. You take the thing away, whatever it is.
So that’s our first job. Think math. Take away the glamor around the word “positive.” In learning theory it just indicates an additive operation. And “negative” indicates a subtractive one. Ergo our first set of definitions:
- Positive means that something is added after a behavior
- Negative means that something is taken away after a behavior
The thing that gets added could be something really great (a bag of cookies), or it could be something awful, like a kick in the stomach. Likewise the thing that gets taken away could be something you’ll miss, like that same bag of cookies, or it could be something that makes you sigh with relief when it stops, like a car alarm.
And that sets us up to talk about these “somethings” as consequences. Now we need to lose the idea that reinforcement is good and punishment is bad. Here is our second set of definitions:
- Reinforcers are behavior-increasing consequences
- Punishers are behavior-decreasing consequences
That sounds really dry, but that’s the way we need to think if we really want to understand this. We need to rid ourselves of the other definitions of these terms that leak into our head from pop culture.
Now let’s combine these four items into definitions of the four possible processes:
- Positive reinforcement: Something is added after a behavior, which results in the behavior happening more often.
- Negative reinforcement: Something is removed after a behavior, which results in the behavior happening more often.
- Positive punishment: Something is added after a behavior, which results in the behavior happening less often.
- Negative punishment: Something is removed after a behavior, which results in the behavior happening less often.
Here is a chart that shows the processes in quadrants. I used a circle, rather than the more traditional rectangle. We can still see that we are looking at four processes that vary on two dimensions (there’s the math again). My artist friend suggested the circle for use in the movie and I thought that was a great example of thinking outside the box….
Good and Bad
Take a look at the definitions of positive reinforcement and positive punishment again. “Something is added after a behavior, which results in the behavior happening more/less often.” Do you see that they differ by only one word? “More” vs “less.” So if increases OR decreases in behavior can happen by adding things, those things must be pretty different if they are achieving the opposite result.
That’s why you will see many definitions of the procedures of operant learning that do describe the consequences as “good and bad” or “desirable and aversive.” They are not wrong, and those definitions can be very user friendly, especially when you are just starting out. Here is a great example of learning theory explained in very lay terms by trainer Leah Roberts. She uses “good” and “bad,” but carefully notes that these consequences are defined by the dog.
That’s the caveat about using the good/bad terminology. We tend to have pretty strong ideas of what constitutes good and bad things, and they may be very different from how the animal responds. It’s the animal who matters, though. For instance, most of us would not classify yelling at a dog as “adding a good thing.” But yelling at a dog can reinforce what they are doing when we intend the opposite. Imagine a dog who has been locked outdoors all day and is now barking at the back door. Nothing happens, and nothing happens, then mom finally sticks her head out and yells, “Stop that barking!” and closes the door again. Social contact for a dog who has been starved for it. Do you think he will stop barking?
Likewise, we humans seem to love patting dogs on the head, and there are whole schools of training that claim that their dogs work for “petting and praise.” Well, I included a head pat in the movie. Here’s a big hint: it’s not in the positive reinforcement section!
Increasing and Decreasing Behavior
We have Skinner to thank for some of the terminology confusion, but also, reinforcement and punishment are intrinsically more tricky to understand than they first appear to be, even when you get the positive and negative straight. If I give my dog a cookie when she sits, and she snarfs it down, have I just reinforced the sit? We don’t know! Part of the definition of reinforcement is that the consequence results in an increase in the behavior. So we don’t know if we have reinforced anything until the future, when we see whether the behavior actually increases.
Punishment is even more confusing because it appears to work right then. The dog barks, you yell at the dog, the dog stops barking for 30 seconds. Have you punished the barking? In the vernacular sense, anybody would say yes. In the Skinnerean sense, again, we don’t know yet. It got interrupted. But punishment has to do with future behavior. We only know if the barking has been punished if the barking decreases.
There are several uses of aversives portrayed in the movie that are very common in real life, done by people with the intent of punishing or negatively reinforcing a behavior. One example is when a dog barks in a crate. Some people keep a spray bottle of water next to the crate and give the dog a spray when he barks or whines. I’ve seen people do this day in and day out, even asking their friends to go by and give the dog a spray as a bonus. Are they punishing the barking? Apparently not. The barking has not decreased. Perhaps the dog likes being sprayed, or more likely in this context, there are other competing consequences of the barking that are reinforcing it.
Did Reinforcement or Punishment Occur?
That’s one more thing to consider: when you see the scenes in the movie, you do not see what other factors may influence the dog in real life. They might be from the environment or from her history. Back to the head pat in the movie: it’s in the positive punishment section. But honestly, Zani’s lying down on cue is not going to decrease because I patted her on the head one time after she did it, even though she clearly doesn’t care for it. She has been reinforced with food and play thousands of times for lying down. The one head pat, although unpleasant for her, will make hardly a ripple. But if getting patted on the head were the only consequence Zani ever got for lying down when I cue it, lying down on cue would decrease, and she would probably be avoiding me to boot.
So the predictions I give for each behavior in the movie (predicting that it will increase or decrease) would most likely be true if the consequence shown were the only consequence for that behavior.
You have probably noticed that I have not used the term “operant conditioning.” This is because of what I have learned from taking Susan Friedman’s wonderful course. One of the things the Applied Behavior Analysis folks focus on is precision in language. So although “operant conditioning” is an old and commonly used name for what I am writing about here, I am taking her lead and referring instead to “operant learning.” Because unlike in classical conditioning, we are not talking about a conditioned response at all. The processes that are commonly shown in quadrants are about learning via consequences.
And ah, quadrants! Susan kindly pointed out to me that there is no official term “The Quadrants” in Applied Behavior Analysis. (Guess what the original name of the movie was!) That’s so sensible, too. Why would you refer to four processes that are the cornerstone of so much of what you study “the sectors” or “the divisions”? We are discussing processes. That leads to another one: she uses “processes” for the organically occurring learning in life. She encourages the use of “procedures” when one is discussing intentional arrangement of consequences (i.e. training) by a human.
My use of correct terminology is a work in process. Obviously, any mistakes or unfortunate wordings in this post are my responsibility alone.
Purposes of the Movie
As I mentioned above, the main purpose of the movie is to give several examples of each of these processes to aid the viewers in developing an understanding of the underlying concepts. I had a couple of other reasons for making it, too. One is to show the rather wide variety of consequences that can be used in each of the four processes. Another is to set me up to write a post about the Humane Hierarchy, which is an aid to making training decisions in order to maximize empowerment of the animal.
And that leads me to what the movie is not: It’s not a training aid. All of these processes are not equally desirable! I am including portrayals of some extremely aversive consequences for educational viewing purposes only, and do not condone those practices. You may notice the green, yellow, and red colors in the chart. That is a hint of what I’ll be presenting in the post about the Humane Hierarchy.
Finally! Here it is.
Thanks for watching! Stay tuned for all sorts of stuff, now that I have finally finished with this one!
- The Humane Hierarchy
- What’s in a Name
- Cricket’s Safe Room
- Arguing with Logic and Grace
- When Management Succeeds
- Level 1 Breakfast (quick behavior drills)