 Hey everyone, it's Veronica Howard. We're going to talk about shaping really briefly. What I'm going to do today is just talk about some of the history of shaping to give you a definition of it, a brief example of it and tell you some places where people typically go wrong when using this procedure. The problem with this particular lesson is that shaping is often something that's much easier to talk about and something that's easier to understand conceptually and much harder to translate into practice. So if you have the opportunity to do it, I would really encourage you to get your hands on trying shaping, actually teaching someone something new and teaching someone something new using only consequences. Don't give instructions, don't give guidance, just see what you can get in terms of using reinforcement because that is going to show you some of the difficulty with using shaping and some of the value that it has. When we're talking about shaping, it's one of the only ways to teach brand new behaviors in a repertoire. So shaping is really important because very often we're working with folks who may not be able to understand us. They may not have receptive language, but we need to have a way of communicating what we want to need from them without actually communicating with words what we want to need from them. For instance, Skinner was challenged by a magazine editor to teach animals to do things that were really complex but in a very short time span. So this is a photo spread from a magazine in the 1950s. I forget the name of which one, but the magazine editor didn't believe Skinner that shaping could be used to teach all these new behaviors that it was as powerful as it really is. And he said, okay, well, I tell you what, why don't you, if behavior analysis is so amazing, why don't you teach a dog to run up a wall? And Skinner said, yeah, I could do that. You know, challenge accepted. So what I'm showing you here is time and how the shaping procedure went. When Skinner started, we're just seeing what the dog will do and within the first minute or so, Skinner is reinforcing the dog whenever they're kind of near the wall. That's what you see at the one minute time. Around four minutes, what Skinner is waiting for is for the dog to really consistently go over to that wall space where the little lines are. Around eight minutes, you see that the dog, when being reinforced, is consistently orienting toward the wall. And then Skinner starts to withhold reinforcement. Now the dog can't earn reinforcement just for being near or looking at the wall. Now the dog has to give you a little bit more. So you see in minute 12 that the dog is actually orienting and putting paws on the wall. And you see from 16 to 20, he's kind of shaping up that repertoire of jumping up the wall. Skinner also did this with one of those little push lever operated trash cans. Let's see if we can get the dog here to open the trash can by pressing that little lever. Now clearly Skinner was not a dog owner because you don't want to give dogs the tools to actually get into the trash, but we'll put that aside for now. Skinner begins by reinforcing the dog whenever they're near the trash can, then withholding reinforcement and reinforcing only when the dog is putting their paw on the little lever. And then once you start seeing that happen pretty consistently, only reinforcing when pressing that lever down. And eventually the dog learns to push down that button and pop open the lid. Now that's a super simple way of describing what I'm talking about in shaping. But again, it's easier to talk about than it is to do. When we're talking about shaping, shaping is the use of differential reinforcement. In a series of successive approximations to a specific target behavior. In those examples, our target behavior was opening the lid of a trash can, it was jumping up a wall. But the way you get there is by using differential reinforcement and being savvy enough, being analytical enough to know when it's time to move on to the next step and when you should wait and stay at the step a little bit more. We use shaping often when we need to teach a new behavior. When we're trying to teach something that's not already in the learner's repertoire. And this can be pretty difficult because we often, if someone knows how to do something, you can just get more of it by reinforcing it. But what if you need a behavior that they've never done before? What if you need to help a learner who doesn't have receptive language, who may not learn from models or in a situation where you can't use simple reinforcement alone and you have to create something new? That's why we use shaping. And the example I want to give you today is actually teaching children language, right? I could use differential reinforcement in a series of successive approximations and many of our parents do. So for instance, when I was younger, maybe my mom is going to get me a glass of milk only when I'm doing certain behaviors and consistently on screen here what I'm going to show you is what will be reinforced, which is on the right. It will also be displayed in green and what will not be reinforced, which will be on the left, which is going to also be displayed in red. So maybe my mom at first will reinforce me only for things like pointing at the refrigerator or pointing at my glass to indicate that I want something to drink. But my mom is never going to reinforce me if I'm crying or screaming or pouting. And then over time, I develop a repertoire where when I want something to drink, I'm going to point at the thing that I want, right? If I'm reinforcing that behavior with milk, I'm going to get more of that behavior. But this isn't language yet. So over time, once I become really fluent, once I'm producing a lot of pointing behavior, maybe now my caregiver can refine that a little bit. Maybe then the caregiver can say, you know what, we're going to provide that reinforcer, but we're only going to provide milk when the client or the learner or the child is pointing at the refrigerator with their glass in the hand, right? So it's like, I want that. And this is where I want you to put it, where they bring that cup to the parent. But now not only do they not reinforce the crying, screaming and pouting, but now they no longer reinforce pointing without a noise. So we've, we refined it a little bit. We've gone a little bit further. We've said, you have to give us just a little bit more than you were giving us before. And then once they get pretty fluent here, we can move forward. Maybe now they need to point at the cup and make a noise like, but we don't reinforce crying. We don't reinforce pointing without noise. And we don't reinforce just bringing the glass to the parent. Now they both have to bring the glass and make a noise. Now they're making the noise pretty regularly. And maybe we say, you know, we're only going to reinforce when they make noises that may have kind of an M sound to it, like meh. And we're no longer going to reinforce any of the stuff from the previous steps. And we're no longer going to reinforce other noises. We're only going to reinforce the cup and the noise. Once that gets pretty fluent, we can then refine. And then maybe we get, we say milk or some approximation of milk that young kids can say like milk. And then we no longer reinforce everything else. So you can see this is going to take time. And this is going to take patience. But you're essentially shaping up a brand new behavior. You're creating something really new from what wasn't there before. But this is so difficult. You guys, it's so, so difficult to do this. And there's a lot of things that I see happen regularly that are problems with different or with shaping. The first big problem is that not reinforcing soon enough. So maybe the target behavior is happening, but the reinforcer is delayed. And so when the reinforcer actually becomes delivered, it can be kind of confusing, like what the reinforcer is for. One of the ways that trainers will often get kind of get around this issue of timing. And when reinforcers are delivered and what's marking or what's communicating, yes, that's the behavior, is they'll use clickers, right? Just a little something that can make a little noise that communicates to the learner. Yes, that was the noise to make. In fact, let me grab my clicker. All right, and back. And I've got it. So just a little something that makes a little noise like and that noise can become a conditioned reinforcer to communicate to the learner. Like that's it. That's the behavior, right? That's the exactly the thing we want you to do. Some folks find this a little bit insulting. They think of this as being dog training as if the procedures that we use with dogs should be different than the procedures that we use with human animals. I don't know if I agree with that. And I'm going to post a podcast for you guys to listen to in the comments and in the study guide. So you can have a look at that. But check out the description, check out the study guide for that link. It's a way that clicker training has been used with folks in advanced medical training, because it's much easier to communicate. Yes, that's right with something like a clicker than it is to actually deliver the reinforcer or deliver a lot of language, not reinforcing soon enough. That's a big problem. It's confusing. It doesn't tell you what behavior you want to see more of. Another problem I see is not reinforcing enough, right? Once the behavior is omitted only once, then the reinforcer is delivered then what? People often don't think, oh, I have to reinforce this like a lot. I have to reinforce this regularly and often. Remember the matching law. If you want to see an alternative response happening, you have to reinforce it a lot. You have to give lots of reinforcement, especially at first when the new response is being learned. Because behavior is not reinforced enough or often enough, it can lead to some emotionally charged or extinction-induced emotional responding. Now we actually want to capitalize a little bit on extinction-induced variability because we want to get some of those new behaviors. We want to get a little bit of new behavior, but we don't want to put a behavior on so much extinction that we get a lot of those emotional side effects. This one is a really tough line to walk. How much extinction is enough extinction? How much extinction is too much extinction? You know that if you're getting into emotional behavior, you've gone too far. It's too much extinction and you may need to go back to a previous approximation, provide a lot of reinforcement, and then move forward. Again, fine line to walk because sometimes I see that people may reinforce or stay at one approximation too long and it can kind of lead to something like some perseverance on that particular step before moving forward. This is a skill that's much easier to talk about than it is to do and I would encourage you if you have the opportunity to try this. If you have another person that you live with who might be willing to let you experiment on them or a pet that you own or someone, I mean heck maybe even someone that doesn't know it's coming, I'm not going to say anything. I won't report into the IRB but give it a try because shaping is really powerful but it's also really challenging and see what what you can come up with. We'll see you guys next time.