This is the first in a series of posts based on my notes from the 2020 Art and Science of Animal Training Conference that was held in Hurst, Texas on February 22-23, 2020. To learn more about the conference, you can visit the conference website.
While I try to take accurate notes, it is possible that there are errors or that some detail is lacking. If you post a comment or email me, I can try to clarify or provide some additional information. Many thanks to the speakers and organizers who allow me to share.
Coercion Without Aversive Stimuli: Dr. Joe Layng
Coercion – to compel to an act or choice.
Coercion often often involves the application of aversive stimuli or physical force. Positive reinforcement is often offered as an answer to coercive practices. But is this always true? In his talk, Dr. Layng spoke about how positive reinforcement can be coercive, how to recognize when it is, and how to avoid it.
How can positive reinforcement be coercive?
A nonlinear analysis* suggests positive reinforcement may be used to compel to an act or choice. There are very few times when you can only do one thing. As soon as you start to limit the options, the more likely you will end up in a situation where coercion occurs.
This idea dates back to the work of Israel Goldiamond in 1976, in which Goldiamond used degrees of freedom as a measure of an individual’s ability to make choices under different conditions (contingencies).
Degrees of freedom is defined as the number of alternatives minus 1 (n-1). If you have three options, your degrees of freedom are 2 (n -1). If you have any degrees of freedom (more than zero), then we call this choice. If you have zero degrees of freedom, then we call it coercion.
Understanding Degrees of Freedom
The degrees of freedom under any contingency can be affected by different factors and changes as they change. Joe shared the following example:
Let’s say you live in an area where there are three options for generating income. Each option requires a set of skills. The possibilities look like this:
In this scenario, you could work in the mine, mill, or field. You have two degrees of freedom (3 – 1 = 2).
But, what if you have no milling skills?
|Mill||No Milling Skills|
Now you can only work in the mine or fields. You have only one degree of freedom (2 – 1 = 1) because the option of working in the mill is not available to you.
And what if there are no fields, then …
|Mill||No Milling Skills|
|No Fields||Field Skills|
You now only have one option – working in the mill. You have zero degrees of freedom (1 – 1 = 0) and you are coerced into working in the mine. You are still getting paid, but you have no choice about how you earn that income.
A dog example
You are spending time with your dog and he can play with a toy or sit to get a bite of food. He has one degree of freedom because there are two options available to him (play or sit to get a treat) and they both lead to reinforcement. Note that this is not about having two types of reinforcers. He can play OR sit.
Now you might be thinking that individuals always have alternatives. In the first example (the mill, mine, field one), the individual doesn’t “have” to work in the mine even if that’s the only option – but maybe he does, if he wants to eat.
To understand this, you have to look at the consequences. On certain occasions, behavior can have more than one consequence. The consequence that governs the contingency is the CRITICAL consequence. If an animal is food deprived, then food becomes the critical consequence and individuals will choose to do behaviors they “don’t like” if that’s the only way to get it.
A pigeon example
A pigeon pecks a key and receives a painful electric shock for each peck. At the 50th peck, it gains 3 seconds of access to food. Because it is deprived, food is the critical consequence and it will continue pecking (despite the shock) in order to gain food. Not only will it keep pecking despite the shock, but if you put the pigeon under conditions where pecking does not produce a shock, it gets upset because no shock = no food. The pigeon has zero degrees of freedom.
Lighted key – peck -> shock -> food
Lighted key – no peck -> starve
But what if you add another way to get food, with no shock?
If you add a treadle that the pigeon can push to get food, it will immediately abandon the lever and press the treadle instead (pigeons are not stupid). The pigeon now has two options (key peck or treadle press) which gives it one degree of freedom (2 – 1 = 1).
In the first case, the pigeon was coerced (by positive reinforcement) into pecking and turning on the shock because it had limited options. You only have genuine choice when the alternatives are equal. If the pigeon is well fed, then food is no longer the critical consequence and it will make different choices.
Once the critical consequence is equivalent for each contingency, the other consequences can have their effect. For example, a well fed pigeon could clean his feathers instead of pecking the key (and getting shocked). When the alternatives all offer equal consequences, then we call this genuine choice.
Back to the mine, mill, field example
Besides income, what are the consequences? The other consequences are black lung disease, grain explosions, exposure to pesticides, distance to work, etc. If all consequences are equal (income is NOT a critical consequence), then the other consequences can now have their effect.
Contingencies that arise as a direct result of the behavior are called program or activity specific consequences and they typically have their effect only when the critical consequences across alternatives are the same.
Increasing the degrees of freedom
Dr. Layng said that recently he has been hearing a lot about start button behaviors and how they provide choice. Do they? Could we add more choice (degrees of freedom)? Here’s what happens with one start button:
|Doesn’t do start button||on cue||no click/food|
The animal has zero degrees of freedom. Depending upon whether the consequence is critical or not (and some other factors), this could be considered coercion.
But what if the animal had two start button behaviors and he could offer either one?
|Start button 1||cue 1/routine 1||click/food|
|Start button 2||cue 2/routine 2||click/food|
In this situation, there is one degree of freedom and the other consequences of the routine can now have their effect. With two routines, you can learn more about the inherent “rewarding” value of each activity. You can pick up on the nuances of behavior by providing alternatives.
Dr. Layng chose a start button behavior as an example, but you could accomplish the same thing by allowing the individual to offer one of several behaviors. In a past ASAT presentation, Ken Ramirez talked about allowing the animal to choose between behaviors as a way to allow the animal to say “no.” You can find my notes on it here.
Beyond animal training
If you have limited degrees of freedom, this often leads to feelings of isolation and can contribute to the formation or joining of cults, terrorist groups, etc. You are compelled to behave a certain way through a series of consequences. Then, once you are in the group, you can be reinforced for striking out at others outside the group and become even more limited in your choices.
Almost all training, even with children, often restricts alternatives. That is, the behavior is coerced by arrangement of alternatives and critical consequences. The question is not whether coercion is used, but what is the rationale for its use and to what degree?
If you are interested in reading more about degrees of freedom and Goldiamond’s work, you may find this article helpful.
*Curious what he meant by “nonlinear analysis?”
I was, and found this information in the book “Social Work Practice: A Critical Thinker’s Guide” by Eileen Gambrill.
If you read this article and felt tempted to throw your hands up in despair because it seems like using positive reinforcement keeps getting more and more complicated, then take heart. This conference is about pushing the edges of what we know and learning to look for the details and nuances that separate good training from great training. Wherever you are in your training journey, there’s always the next step, and that’s a good thing as it means we all have the opportunity to keep learning.
Because this is the first presentation I’ve shared, you don’t get to read it in the context of the entire conference. I think once I’ve shared some of the other presentations, you’ll see that the conference had several themes. One was about looking at the big picture and seeing behavior and consequences in context. I’ll talk about this when I write up my notes on some of the other presentations, particularly the keynote and Mary Hunter’s talk.
Another theme was about how important it is to re-evaluate what we know on a regular basis. Dr. Rosales-Ruiz talked about how it was good to look at the science again every 10 years and ask ourselves if we need to update our training practices or how we think about behavior. In his presentation on CAT, he described the original procedure and how it has changed to the current version of CAT.
I think this talk was an important one because trainers who are new to positive reinforcement often think that providing positive reinforcers (as measured by their effect) is enough to guarantee that the horse has choice. But that’s not always the case. I’ve been training for long enough to have encountered situations where I was able to use positive reinforcement to “coerce” a horse into doing something that he was not comfortable doing. In some cases, the reluctance was visible from the beginning and I adjusted. But, in others the horse seemed very enthusiastic about the training and it was not until other changes occurred that I realized I needed to stop and re-evaluate.
I think out best defense against unknowingly coercing our horses is to be aware that it can happen, monitor the horse’s body language, and provide opportunities for the horse to express his preferences. Off the top of my head, here are a few suggestions for how to do that:
- If you train with food, make sure the horse is not hungry and/or have other food available in the environment.
- Pay attention to the horse’s body language through the whole training loop including cueing, doing the behavior, and getting the reinforcer. How a horse takes a treat is valuable information. You can read more about this here.
- If I’m teaching a new behavior, or asking for a behavior that is more difficult, I practice it in very short sessions, taking breaks or alternating with other easier behaviors. Doing this makes it easier to tell when my horse is getting tired or reluctant to do particular behaviors.
- Set up some more “flexible” training sessions where the horse can choose which behaviors he wants to do. This may require some advance thought about stimulus control and how to set it up so the horse knows he knows he can choose. I’ve done this a little bit with behaviors that have object cues but it’s something I would like to play with more.
- Teach your horse enough behaviors that he has many different ways to earn reinforcement. If possible, allow him to offer some of them as default behaviors so he can use them to communicate with you.
- Observe, observe, observe… Did I say observe? Pay attention to things like the horse’s interest in training. Does he want to come out and train? Is he the same in all locations? Does he respond to his cues promptly? To all cues promptly?
I love that we are all becoming more aware that animals will perform behaviors for positive reinforcers if that’s the only way to get access to them, and that its our responsibility to make sure they are willing participants. This talk added another layer to my understanding of how to do that and it certainly gave me some food for thought. I’ll finish with a final thought from Dr. Layng. I can’t guarantee this is a direct quote as I was scrambling for my pen, but it’s close enough.
It’s not about making the activity fun, it’s about finding the fun in the activity.
I’ve been blogging about the ASAT conference for several years now. You can find a list of them in the articles page on this blog, but some older ones are on my website, www.equineclickertraining.com.
If you are interested in learning more about how to clicker train your horse, check out my book, Teaching Horses with Positive Reinforcement, available on Amazon.