This the second in a series of posts based on my notes from the 2018 Art and Science of Animal Training Conference that was held in Irving, Texas on March 24-25, 2018.
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. To learn more about the conference, you can visit the conference website.
Dr. Jesús Rosales-Ruiz gave a short (20 min) talk on movement cycles . This was part of a series of talks on how to build precise behaviors by understanding and analyzing movement.
Jesús started by sharing B. F. Skinner’s definition of behavior (1938)
What is behavior?
“The movement of an organism or its parts in a frame of reference provided by the organism itself or by various external objects or fields of focus.”
The definition includes two components:
- What part of the organism was involved(movement)
- How it relates to the environment (space and time)
Examples: Pressing (movement) a lever (environment) or looking towards (movement) the light (environment)
It is convenient to speak of this as the action of the organism on the outside world, but it is sometimes easier to deal with an effect than with the movement itself. In the case of the production of sounds, we can’t see the vocal cords moving so we have to rely on some other movement, such as the movement of the lips or mouth. Therefore, when defining behavior, we need to be aware of both the observable behavior and the physical action that creates it.
Ogden Lindsley, who studied under B.F. Skinner, introduced the idea of “movement cycles” in 1969. He said that it was not enough to look at an isolated behavior. Instead we should study the entire cycle which contains the behavior.
- Each response has a beginning and an end.
- The behavior is not done until the organism is in a position to do a new one.
- Sitting – we usually define sitting as contact with the chair, but the cycle starts when you are standing up, includes all the steps that immediately precede sitting down, sitting, and then all the steps that follow – until you are in a position from which you could sit down again.
How do movement cycles relate to shaping?
Jesús had a slide with a graphic of a movement cycle. The picture at the beginning of this post shows something similar, with the movement cycle for a dog sitting. The cycle starts at 9:00, the middle is at 3:00, and the end is back at 9:00 again. Between 9:00 and 3:00, there would be all the steps a dog goes through as preparation for sitting. Between 3:00 and 9:00, there would be all the steps a dog goes through as he stands up. In many cases, the behavior in the middle of the cycle is the one that the trainer would click.
He had another graphic of a chain as a series of movement cycles that were linked together.
How can we use the idea of movement cycles in shaping?
We can focus on the process of getting the behavior, instead of the outcome.
- Begin shaping at the beginning of the movement cycle.
- Follow the movement cycle as you shape.
- You can feed to produce the beginning of the movement cycles, then click the action in the middle of the cycle.
Kay Laurence shaping Quiz to put her foot on the dice, showing the difference between clicking for touching the dice with her foot (the middle of movement cycle) and clicking for lifting the leg (the beginning of movement cycle). Quiz learned the behavior more quickly and with fewer errors when Kay clicked for lifting the leg and added in touching the dice later.
Alexandra Kurland’s students using microshaping to teach a horse to step back, showing how you can shape a step back by clicking for a tiny weight shift (starting at the beginning of the cycle) and then change the timing so the click marks a behavior that is farther into the movement cycle to get a full step back.
Mary Hunter shaping Ginger to go out, touch a stool and return. The video showed what happened when they moved the click later and later in the movement cycle. When you reinforce, you reinforce the whole movement cycle, not just the part where you click.
- Teach going out and touching a stool. Click for stool touch (1/2 way through the cycle)
- Then move the click to ¾ of the way through the cycle. Now she is clicking as the dog comes back after touching the stool.
- This is moving the click in the direction the behavior is going.
- When you move the click later in the cycle, you can do hundreds of repetitions before you see any deterioration in the behavior. I don’t think they actually did this – he was saying it as a general observation – but someone in the audience pointed out that the behavior was already changing as soon as he moved the click. Perhaps more data is needed…
Mary Hunter with Drill Bit (dog – unusual name!) and clicking for attention. Mary is sitting and Drill Bit is lying down watching her.
- She is clicking for attention.
- When she starts to shape for more duration, several unwanted “extra” behaviors start to creep in (tail and leg movement).
- Go back to the beginning and shape in small increments to clean up the behavior.
When we teach and analyze behaviors, the tendency is to focus on a specific moment in time, the moment when the desired behavior happens. But, it can sometimes be more effective to look at the entire movement cycle, which includes what happens both before and after the behavior occurs.
- Focus on Movement, not Outcome
- Use movement cycles to define or plan the shaping steps
- Use movement cycles to clean up behaviors
A personal note:
A few years ago when I was teaching hoof handling to my young horse, I became more aware of the importance of looking at the entire movement cycle in order to avoid reinforcing unwanted behaviors between the click and treat. She arrived with the habit of striking when her front feet were handled. I could shape a nice leg lift, but as soon as I clicked, she would strike out and then put her foot down. Not what I wanted.
So, I changed my shaping plan and I taught her to pick up her foot a tiny bit, and then allow me to place it back down. To do this, I started by clicking for a foot lift as normal, but then, instead of building duration for holding it up, I mixed in some clicks for allowing me to put the foot back down. When she could pick up and put her foot down nicely, I slowly added more to the middle of the behavior until I could pick the foot up and hold it up for longer and longer periods of time. By paying attention to the entire cycle from the very beginning, I was able to avoid reinforcing unwanted behavior between the click and treat.