Alright, so now that we’ve gone over reinforcement, punishment, shaping, and extinction, we can move onto the next step in our behavior journey: Stimulus Control and Discrimination.
Remember back in reinforcement, we talked about how behaviors are usually controlled by the consequences of what happened to them. If something the subject likes happens, then the behavior is more likely to happen again (reinforcement). In comparison, if something the subject doesn’t like occurs, then the behavior is less likely to happen (punishment).
Stimulus Control is the idea that a behavior will occur in the presence of one stimulus, but the behavior will not occur in the presence of another stimulus. Compared to what we talked about earlier, stimulus control works with what happens before the behavior – what we’re going to call antecedents. These antecedents help our subjects figure out if they’re going to get reinforced or not.
Discrimination is just knowing when the presence of a stimulus will result in reinforcement and when it will not. Before you think it, no this has nothing to do with any forms of discrimination in society. Discrimination in this context is taken as just figuring out which stimuli will give you a reinforcer. I also lightly touched on discrimination in my video games and violence article, but we’re going to take a closer look at how discrimination works.
Generalization is the opposite of discrimination. It’s when your subject takes a behavior and applies to more situations than the training one.
Before I get to an animal-specific example, let’s take a look at a situation where you’d see this in everyday life.
Let’s say you were put behind the wheel of a car and also completely forgot how traffic lights work. Your job is to drive from wherever you are back to your home. On your wondrous journey, you come across a traffic light. The traffic light is already showing a green light. Not knowing what that is, and wondering why an odd Christmas light would be shining in your face, you continue to press the gas and keep going on your merry way. In this case, getting the car closer toward your destination would be the reinforcer (or at least, not being intruded by anything).
However, let’s say while you were on your merry way, you reach another traffic light. But this time, the light is red. If you press the gas now, oncoming traffic can honk, hit, maim, or wreck you for moving into their way. In fact, the honking part is exactly what happens. All of these factors would be punishing your behavior of pressing the gas pedal.
“But hold on” you would say. “Just a second ago, pressing the gas was okay, why is it wrong now?”
The answer here is that the antecedents of the lights were telling you when pressing the gas would be reinforcing and when pressing the brakes would be reinforcing.
This is exactly how stimulus control works. One antecedent would mean that you’re going to receive a reinforcer at the end of it, while another antecedent would mean you’re about to get punished, or nothing will happen.
For animals and animal training, this concept is even more important.
Animals and Stimulus Control
Every time you’re giving a command or a hand signal to an animal, you’re practicing stimulus control.
In fact, as trainers, there’s a reason why we do these strange little hand signals or add vocal commands. Yes, it helps us communicate what we want to the animal, and the animal knows “oh damn, if I do this behavior now I’m going to get a treat!” But what’s more important is that we’re also telling the animal when we’re not doing the hand signal or command, that they will not get any reinforcers.
Let’s say I’ve taught a dog how to sit already and gave him the hand signal associated with sitting. That hand signal works in the same way as the green light in our earlier example – if the dog sits then they know everything is all good, and in fact, the dog will get exactly what they want out of it (a treat).
However, if you’ve owned a dog before and you just taught a new trick to your dog then you know what happens next: they start doing that trick all. the. time.
Were you going to go to the kitchen? Well, first you have to see me sit. See? Look I’m sitting. Give me the food. Oh, you’re going to the bathroom? Cool cool, but watch me sit on your feet. See? Sitting. Now give me the treat.
This is where your dog has generalized the idea that sitting will always get him a treat. The world is his antecedent, and he’ll be damned if he doesn’t get his treat. What we need to teach here is discrimination – meaning, the times where it’s okay to sit and times where it’s not okay to sit. This is where antecedents become useful.
If your dog is driving you crazy with out-of-place behaviors cause they want a treat from you, then you have to start pairing behaviors with antecedents. When you’re doing a training session with your dog, say “sit.” Ideally you say this before your dog sits or even as your dog is sitting. You should click and reinforce during this time. Have your dog get into another position where they aren’t sitting (you can walk away so they follow or you can have them do another trick), but after that, say sit and reinforce for doing that behavior. While you’re doing this, make sure you never give a treat unless you said sit first, and your dog sits. If your dog is expecting what you’re going to say and does the behavior before you say or give an antecedent, then do not give them a treat (timing can be a bit fuzzy here, but usually you don’t want to give a reinforcer until you say sit first).
What’s important here is that you’re teaching your dog that the only time they will get a treat for sitting is if you either say “sit” or give them a hand signal. Barking and bathroom problems can also be solved with this technique. For some dogs, teaching them when a bark will result in a pet and a compliment or just a treat can help curb barking problems.
Alternatively, I’ve used the idea of Stimulus Control to potty train dogs. The way it works is to say the command you wish to use for bathroom (“potty”, “bathroom time”, whatever works for you) every time your dog is, well, doing their business. If you’ve done this enough times, your dog will start to associate the word you’ve used with the act of going to the bathroom and will understand what you want. As a result, the command you’ve created can now be used as an antecedent to when it is okay for your dog to go to the bathroom.
How to Read this in SCIENCE
If you’re going to be proactive about learning behavior science, then it would be good for you to know what a discriminative stimulus is.
A discriminative stimulus is the antecedent stimulus that means the subject will be reinforced when it is present. In understandable terms, it means when you do or show this thing, your subject is going to get what they like. For example, you say “sit” and your dog sits, then you give them a treat. In this situation, since you reinforced the behavior of sitting, saying “sit” is the discriminative stimulus.
In scientific literature, a discriminative stimulus will be abbreviated as SD.
The opposite of the SD will be the S𝛥 (or Delta). This means it is a stimulus that is not being reinforced. In that same example, if you say “potato” and your dog sits, you won’t give him a treat. In this case, since you didn’t reinforce sitting after the word “potato”, the word “potato” is an S𝛥.
Scientific experiments will also use the symbols SR and SP. SR means a reinforcing stimulus, while SP means a punishing stimulus. Usually, you’ll see it in behavior chains that can look like this:
SD -> R -> SR
R just means the response.
So that chain would mean the subject was shown a discriminative stimulus (like a hand signal for “sit”). They then gave a response (like sitting), followed by a reinforcing stimulus (a treat).
If we head back to the traffic light example, the green light would be the SD, the R would be continuing to hold the gas pedal, and the SR would be going along on your journey unimpeded.
Stimulus control is the difference between your dog running around barking non-stop because you taught him how to speak, and your dog sitting patiently waiting for your signal to bark.
While it sounds complicated at first, it’s just like the rest of dog training and Operant Conditioning – the more you practice at it, the better you’ll get. A lot of stimulus control is going to revolve around you keeping an eye on what your subject is doing and how to make sure you control the antecedents that go before a behavior. Once you start pairing behaviors with antecedents, then you’ll find it will be easier and easier to communicate with your dog.
[…] next few articles will talk about Antecedents a bit more. I’ve covered antecedents a bit in my Stimulus Control article, but It’s important to understand that those concepts should never be confused with Classical […]
[…] briefly also talked about this in my Top 5 Motivation article and I talked about Antecedents in my Stimulus Control article, but I wanted to spend some time to talk a bit more in-depth about it […]