Dr. Jesús Rosales-Ruiz is an associate professor at the University of North Texas in the Department of Behavior Analysis and serves as chair of the department. At ASAT, he always shares information about the science behind many of the common procedures and practices done by teachers of all species. At this conference, he gave a short presentation on the topic of context cues.
What are context cues?
The phrase “context cues” is thrown around by many people, but not often clearly defined. Let’s look at what it means.
- Defined by Catania (2014)
- The consistent features of a situation, such as the chamber in which an operant session occurs.
- Experimental contexts acquire behavioral function because they are embedded in still larger contexts that include the experimental session.
This is a good starting point, but it’s not that useful. Instead, it may be better to talk about Constant Stimuli.
Constant Stimuli (SSC):
- Defined by Goldiamond and Thompson (1968)
- Stimulus present during both extinction and reinforcement
- We can label all the stimuli that are present (room light, room sound, round disk, room heat, etc.) as constant stimuli. They are different than the discriminative stimulus (SD), a red disk that tells the pigeon when to peck at the disk.
Behavioral Function of Constant Stimuli
Note: SSC = constant stimuli, SD = discriminative stimulus (cue), S-delta = stimulus that indicates reinforcement for a behavior is no longer available
- SSC do not control behavior directly as the SD and S-delta do
- Changes in the SSC disrupt the control of SD and S-delta
- The disruption is attributed to stimulus change, or novelty
- The disruption is usually called a distraction
- Contingency control (response to SD) is reestablished with repeated stimulus change
For example: A behavior may deteriorate (quality, response to cue, etc.) when practiced in a new location where a train passes near the training area. But, if the animal is regularly trained in that location – and the train continues to pass by during training sessions – the animal will learn to ignore it. The change in the constant stimuli (passing of train) disrupts the behavior until it becomes part of the constant stimuli.
If contingency control is not reestablished (the animal continues to react to the train), then a program may be implemented to control for such disruptions. This process of asking for behaviors under different conditions is considered a form of proofing. Is this a good way of proofing? Maybe not… (note: this was Jesús’s question, but he didn’t expand upon it)
What are the antecedent conditions?
One way to think about the effect of constant stimuli (SSC) is by recognizing the relationship between them, the discriminative stimuli/cue (SD), and the response (R).
- If reinforcement occurs in the presence of SSC – then, the
SSC may produce control because they create antecedent conditions that indicate reinforcement is available.
- You could write this as:
- SD -> R (cue leads to response/control over behavior)
- SSC -> SD -> R (cue and context acquire control over behavior)
- SSC -> R (context controls behavior). A classic example is the dog that starts throwing behavior as soon as you enter the training room. The context (training room) now becomes a context cue for the dog to start offering behavior (it leads to a predictable response).
Examples of the Effect of Context Cues
Jesús had a number of examples that showed how stimulus change (a change in the constant stimuli) could affect how an animal responded to other stimuli in the environment. In this section, he seemed to use the terms constant stimuli and context cues somewhat interchangeably.
Example #1 : The dog has been taught to let go of a tug toy when the student says “let go,” but the same cue doesn’t work for the mother. The dog has learned the behavior under certain stimulus conditions, but not others.
Example #2: Disruption of cue control due to stimulus change (a change in the SSC )
- Kay has taught the dog to go around a cone. She has trained in the same location and the SSC have been fairly consistent. The SSC have become context cues.
- If she changes the picture (SSC) by holding a broom, the behavior breaks down – the dog hesitates instead of going around the cone.
- This is one reason it’s important to systematically vary any possible context cues, otherwise it’s easy for them to become disruptions.
- The strength of a behavior is not just due to reinforcement, it is also a function of how the animal responds to a change in stimulus conditions.
Example #3: Context sensitive behavior
- The dog shows different behavior under two different conditions – inside vs. outside.
- The dog hides when a specific individual enters the house.
- The dog interacts happily with the same individual when they are both outside.
- The dog’s behavior is influenced by the context in which the individual appears.
Example #4: Dog becomes aggressive in the presence of one individual
- Sabrina (the dog) is relaxed on leash in the presence of just the husband.
- She becomes aggressive when another person is present
Example #5 : Contextual Control of Reinforcers
- Fever – 8 year old border collie
- She was trained at an early age to switch between food and toy reinforcers, even within a session.
- She happily switches between accepting the two types of reinforcers mid session and in new locations.
- As an experiment, they trained a new behavior where the only reinforcer was food. She’s retired and doesn’t do much training, so she wasn’t doing other training sessions that mixed food and toys.
- They set up the environment so there were specific context cues:
- It was a new training area (Fever is not usually allowed in the workshop)
- She’s never on leash in the house
- She’s rarely trained without other dogs present.
- She learned the new behavior and performed it well for food reinforcers, but when she was offered a toy reinforcer, she would not accept it.
Jesús talked a lot about how context cues can become reliable predictors of what behavior an animal will offer, and how stimulus change (disrupting the constant stimuli) can affect our training. But, he didn’t talk much about how context cues can be used to our advantage. So, I thought I would just throw out the idea that if you understand context cues, they can be very useful.
For example, my horses have learned to offer different kinds of behaviors in different environments. They know to stand still in the grooming stall and wait while I walk around them, but they might not do the same thing if I ask them to stand in the aisle. In the aisle, they are more likely to walk with me if I move away from them. I can certainly use cues to tell the horse to wait or come, but it’s actually more convenient for me if they learn to offer different different behaviors in each location.
Here are some other uses and examples
I often teach different behaviors in different locations. When I am doing ground or ridden work, we often work on circles in this part of the ring, shoulder-in in another part, and haunches-in somewhere else. This can eliminate a lot of confusion in the early learning stages. With dogs, I might train one behavior in one room and another in a different room.
I can use different objects as context cues. As I learned from Kay Laurence, the objects don’t have to be part of the behavior. They can function as context cues just by being present in the environment. The rug that is used when I practice backing, the table that the dog sits on when I practice downs.
I can use different context cues to indicate different types of sessions. It can be confusing to an animal if she doesn’t know whether or not to offer behavior at the beginning of a training session. So, a lot of trainers have context cues that tell the animal when they are doing a shaping session, and the animal should offer behavior, vs. when they are doing a cued session and the animal should wait for the cue.
I can use context cues to provide a basic level of stimulus control for behaviors that could be unsafe.
- When I taught Willy and Rosie to do Spanish Walk, I wanted to be able to control when they offered it, right from the beginning. I was able to do this by putting a lot of context cues in place. I only trained at night, in the aisle, with a specific target stick, and with one type of reinforcer. Both horses very quickly figured out that it wasn’t worth offering leg lifts at any other time.
- I wanted to each Aurora to trot on a lead, but I didn’t want her to offer it at other times. To create some stimulus control over the behavior, I had a specific configuration of poles and cones in my ring and we only worked on trotting when the arena was set up that way.
- I taught Rosie to do body part targeting in her stall and I only worked on hip targeting (moving the hip toward me) after we had done ear, eye, and jaw targeting. She learned that hip targeting would not be reinforced unless we were in the context of a body targeting session.
The big thing that I learned about using context cues is that they are always out there, so you need to pay attention to them and decide if you want to use use them or not. It’s very easy to get in trouble if you are not paying attention to some stimulus in the environment and it turns out that your animal considers it to be important information. In her talk, “Sculpting the Cue,” Emily Larlham offered some tips of avoiding this kind of issue.
While I try to be accurate in my note taking, there may be some errors either in my understanding or my presentation of the information from this talk. If you have questions, feel free to contact me or leave a comment.
Thanks to Dr. Jesús Jesus Rosales-Ruiz for allowing me to share his presentation. Thanks to the ORCA students and to the organizers of the Art and Science of Animal Training Conference for all their hard work on putting on this great event.