Ann and Panda

Ann Edie: Teacher and Student of Operant Conditioning Part 2: Putting Behavior Science to Work: Behavior Science by and for the Disabled

Ann Edie was born legally blind.  The culture of the 1950s drove her parents to try to help her resist disability; if she could see at all, she should try to develop her vision to avoid being labeled as a disabled person.  It’s easy to see why, when you consider the everyday challenges faced by people with disabilities. Even doctors at that time recommended this approach. Ann was prescribed very thick glasses and attended “sight saving” classes, even though classes were available for blind students to learn Braille and other techniques to help them read.  Ann’s classroom had strong light and they used large-print books, and she did learn to read with her face almost directly on the page under a very bright light. She learned mostly through the teacher’s one-to-one instruction and became a very good listener. Her mom helped her and she used recorded books in high school. Ann managed to get through her early schooling and most of college without learning Braille.  

By the time Ann reached college, she had learned how to persevere to meet her goals.  In college, she carefully listened in class and used recorded books and live readers she hired or that the state commission for the blind provided.  She could read her notes when she used a dark pen and wrote large letters. She was a good note-taker, to the point where friends in study groups borrowed her notes.  By the time she finished her third degree, she found an opportunity to get Braille instruction manuals from the commission for the blind and taught herself to read and write it.  She learned how to use a white cane to navigate through the world, happy to find these helpful aids available to make her daily tasks simpler and allow her to excel in her studies to become a teacher. 

As an animal lover, Ann was actively learning to train her own dogs for competition in shows.  Ann had always known about guide dogs for blind people and as someone who enjoyed having animals in her life, she really wanted to get one.  

Ann’s first guide dog, Bailey, was her partner from 1991 and for nine years until his death.  Bailey had been trained to a high standard, but the techniques the program had used were not primarily positive reinforcement.  The tradition is to use collar corrections, or punishment, to teach dogs certain behaviors are not to be engaged in. Still, Bailey responded beautifully to positive reinforcement training for maintaining his skills with Ann, starting as soon as they graduated from the program and for the rest of his life.  

After Bailey’s death, Ann got a guide dog from a different program.  Even during the initial orientation this dog would bark and pull towards other dogs.  Ann found herself trying to work with a dog that would lunge at other dogs, cats, squirrels, really anything that moved, including light reflections.  Trainers from the school came to Ann’s home to teach her how to punish the dog to stop the undesired behavior, since this was the training technique the program used, even though they used some food treats in their training program. 

Ann learned how to administer the punishment, although she was conflicted about the practice.  By this time, Ann had become a teacher, discovered some of the power of positive reinforcement through training her dogs, and had learned how dogs learn.  She knew the importance of associations, especially between scary or threatening events and painful input. Ann knew her guide dog legally belonged to the training program at this time and she was caught between her own values and the integrity of following the requirements of their contract stating she must train the dog using the methods prescribed by the program.  The dog’s reactivity increased with each incident and the program ultimately retired her to a home where she was a pet rather than a working guide dog. As a student of behavior science, Ann had learned about the marriage between operant and classical conditioning and how visible behavior and emotional response were always happening at the same time. The job of a guide dog requires full focus on the trained tasks of responding to the blind handler without distractions of fear or concern about whether another dog is nearby.    

At this point, Ann brought her third guide dog home from the same program, hoping for a dog more like Bailey and less like her second, reactive guide dog.   Unfortunately, this third dog was also reactive toward other animals, though at first it seemed like the response wasn’t too bad and might be able to be changed so the dog could continue to work as Ann’s guide.  Active in training her horses by this time, Ann knew she could start to change the dog’s behavior as well as his emotional response to other dogs he saw when out on a leash. She worked with the program trainers sent to help her by using their prescribed methods of punishing the behavior they didn’t want.  At the same time, she worked to condition with positive reinforcement to help the dog relax when other dogs were around. One problem was that she could not see when another dog came into her dog’s view, so she couldn’t predict and prepare to start training at the first moment of stimulation. She needed a helper who could alert her to the moment when she needed to begin her dog’s lesson, right when the other dog became visible.  Another problem was that combining two different training methods can cause conflict in the dog. When an animal is punished for the wrong response to a stimulus and given something good for the right response to the same stimulus, it is conflicted, learning to hesitate and consider the options when the stimulus appears, making sure to avoid the punishment. Avoiding punishment can become more important to the dog than getting the reinforcer and learning slows down dramatically.1

Ann attended a training conference where Bob and Marian Bailey were teaching a mini Operant Conditioning Workshop where students trained chickens to peck a spot of a certain color among choices of three different colors.  This was a challenge for a blind person, but remember that many tasks seem challenging for blind people and yet they get those jobs done. Bob and Marian had been teaching students operant conditioning through chicken training for many years.  Chickens are good models for animal behavior and served students well in their workshops. The Baileys had taught blind trainers before and knew they had to set up the task so the student could use another sense besides vision to know when the chicken was doing the right behavior, the one they wanted to reinforce.  They helped Ann use her hearing to do the job.  

Sighted students in the Baileys’ Operant Conditioning Workshops place the colored targets they want the chickens to peck directly on the table where they train them. Bob and Marian found a few empty tuna cans, cleaned them up and put them on Ann’s training table.  They placed the colored targets Ann’s chicken was to peck on top of the cans.  

The specific assigned task in the Baileys’ workshops was to train a chicken to peck the target of a particular color and not to peck the other two targets of other colors.  The student gave the chicken a bit of food to reinforce pecking the correct target.  Giving the chicken the food directly over the target of the right color is key to the training, so the trainer must know exactly where the right and wrong targets are.  With vision, this is pretty simple, but to complicate things further for the trainer, an assistant is constantly switching the locations of the targets! This is done not so much to challenge the trainer, but to prevent errors in training the chickens.  Placing the correct target in many different locations helps the chicken learn that it’s not a particular location on the table they should focus on, but the color of the target.  

The process of training a chicken to peck only a blue target among red, yellow and blue targets is complex for all trainers and it may be hard to imagine how a blind person can do it.  But Ann was able to tell when the chicken pecked the correct target as long as her assistant kept her informed of which tuna can supported the correct, or “hot” target, every time it was moved.  Ann memorized the numbers of the tuna can locations, 1, 2 and 3, and she could hear the difference in pitch and distance of the sound when the chicken pecked a can. She successfully trained her chicken to peck the “hot” target every time and to resist pecking the other targets, using her sense of hearing to determine where each target was.  To sighted people watching, it may have looked like magic; how can a blind person know what color target a chicken is pecking? But to Ann, it was business as usual; use the senses that work well. Think about how what you see may not be what you think it is: the chicken was choosing the target by color and Ann was choosing the target by the sound of the location with each peck, using additional data given by her assistant.

When Ann trains her horses, the process is similar.  She can’t see when a horse changes gait or lead, but she can hear the difference.  In partnership with a horse she’s working with, she knows which direction the horse is going.  Using a lead rope helps her tell what the horse is doing also, even though the lead is loose. With her hands, she can touch a horse’s neck, withers or face and feel muscle movement, so she knows the first hint of when behaviors start, when emotional response is changing.  She can hear the horse’s breathing patterns. When using a target for the horse to touch, she chooses something that makes a sound when touched, perhaps with a bell on it. Having an observer when she’s training helps her make sure the behaviors she’s looking for are happening as she’s defined them; observers help all trainers, whether they are blind or not.  Ann can shape the smallest behaviors. Her timing of reinforcement is stellar because she practices regularly. 

While still working with that third guide dog, Ann and Alexandra Kurland, who taught her to train horses, made a plan to train a miniature horse as a guide animal for Ann.   They would use positive reinforcement, which they both knew was their choice for training an animal to do behaviors reliably and impeccably. They acquired Panda, a 9-month-old miniature horse, and began training her while Ann was still working with that third guide dog.  They hoped this setup would allow them to take their time to get Panda’s training just right. That third dog needed to be taken out of guide dog work and be re-taught from the ground up if he was to be able to function as a guide. Because of poisoned cues1 and bad associations that were already part of his make-up, this was not a realistic task and the program retired the dog from guide dog work because of his reactivity.  Now all the eggs were in the Panda basket! It worked out really well – see our story on Ann and Panda

It has been written that blind people can’t use positive reinforcement to train animals because they can’t know when the animal does the desired behavior, but this idea is clearly false.  As a person with vision, you may have to stretch your imagination to understand what blind people can do when given the opportunity. They have to approach a task from a completely different angle, and they have experience creating a plan using other senses to get the job done. Watch Ann working with horses in our film and make your own conclusions about the power of operant conditioning, the power of positive reinforcement training for any trainer who learns the skills and for the animals they train.      

1 Current information on the topic of mixing punishment with positive reinforcement can be found by searching “Poisoned Cue.”  Pioneers of behavior science found that events causing fear, threat, or arousal for animals in training got in the way of a steady process of learning new skills.  Keller and Marian Breland’s book, Animal Behavior, documents this interaction between classical and operant conditioning.

Share this post