This post is the final part of my notes about a horse clicker training clinic that I audited with Alexandra Kurland in Texas in February. Click here for parts one and two, which discuss balance and cues.
“Are you having fun yet?”
This third, and final post of my clinic notes is a bit of a “catch all” post. It covers work and play, using negative reinforcement when clicker training, and shaping games.
Clicker training should be fun. And it is most of the times. Still sometimes, clicker trainers struggle to find ways to make routine or complex tasks fun. If you ever find yourself (or your horse) feeling bored, unhappy, frustrated, or discouraged, try taking a step back and assessing what’s going on. It might be time for a break, a change of pace, or a new approach. I find this is especially important for horses and riders who started out from a more traditional training approach. Many horses and riders have poisoned cues associated with certain activities or have learned that certain things just aren’t fun.
Alexandra Kurland’s work beautifully combines clicker training with dressage concepts to create balanced performance horses with stunning movement. However, even after discovering the fun of clicker training, I know of many horses and riders still have negative associations with arena work and performance work. It can be hard to keep this kind of work fun and interesting for yourself and your horses if you’ve been taught in the past that this kind of work is suppose to be difficult or demanding. One thing that Alexandra told us during the clinic was that everything we do with our horses should be play. We should find ways to make our training be all about play, even when we are doing “work.” Try thinking about all of your training with your horse as playful experimenting, not as demanding, controlled work.
Here’s one fun, but very useful thing that we did at the clinic. Several of the clinic participants played around with walking in sync with their horse. The trainer and horse would start by walking around a large circle of cones. The person’s goal was to try to walk in step with the horse’s front feet. This is actually more complex than it sounds! The trainer has to focus on keeping herself and the horse moving around the circle, while at the same time watching the horse’s feet.
This is a good exercise for getting in rhythm with your horse. Alexandra suggested watching the horse’s knees, not his feet. The person should also think about to lifting her legs from the hips, not from the feet. The horses seemed to really enjoy this “work.” Several were paying attention and watching the handler’s feet intently. Try this out and I think you’ll find yourself walking differently after practicing walking in sync with your horse.
We can make training sessions more fun is by adding variety to the session or by mixing in some of the horse’s favorite behaviors with the behaviors we want to work on. One participant focused on some on in-hand work and also on improving her horse’s Spanish walk. Both of these exercises took a lot of concentration. So, she would frequently take short breaks to play with the toys in the arena, including pedestals, targets, and even a giant soccer ball.
When training, are you having fun? Is your horse having fun? If there are times when you are not, what can you do to make your training more playful for you and your horse? I’d love to hear comments regarding how you have found ways to creatively work on exercises that would normally be considered not quite as fun.
Is it possible to use negative reinforcement during clicker training and still have a happy horse?
I know a lot of horse clicker trainers who struggle or who have struggled with this question. Most horse trainers don’t start out as clicker trainers, but with traditional horse training methods, which rely heavily on negative reinforcement and punishment. Negative reinforcement is often called pressure and release by horse trainers. The trainer applies some form of pressure, such as pulling on the rein, squeezing the horse with her legs, or waving a whip at the horse. As soon as the horse moves in the desired direction, the pressure is removed.
Many horse clicker trainers start clicker training because they’ve seen some of the harmful side effects of negative reinforcement training. They’ve witnessed depressed, shut-down horses who don’t want to perform or who rebel if the rider asks for too much. These aren’t fun horses to work with.
Horses love clicker training because they are working for something, rather than just working to avoid pressure. However, after starting clicker training, many horse trainers wonder how much (if any) negative reinforcement is okay to use during training.
I know some trainers who go to the extreme and try to remove any and all negative reinforcement and pressure from their training. Others, such as Alexandra Kurland, believe that this is probably impractical and impossible. Instead, these clicker trainers try to find subtle and gentle ways to use small amounts of negative reinforcement as meaningful information, rather than as something painful to avoid.
I have a friend, Dolores, who often gives the example of dancing. If you watch two graceful dancers, they are perfectly in tune with each other’s body movements. Their communication and cues are often so light that these cues are almost impossible to see. But, a lot of these are pressure cues. A dancer might gently press on his partner’s shoulder to indicate for her to move another step over or might press in a slightly different way to ask her to shift her weight back a bit.
This is pressure as information and what we should strive for with our horses. The pressure cue is information only, not coercion, force or threat. If you closely follow Alexandra Kurland’s work, you’ll see that she teaches ways to develop pressure cues that create happy horses that are light and responsive.
Clicker training gives you tools for creating these information pressure cues. With shaping, a trainer can start with a light, barely there cue and accept any small approximations or inclinations toward the final behavior. This can be gradually built up into the final behavior. These types of pressure cues are safe cues because they are both light and predictable. There is never any threat of escalation of pressure or punishment for wrong responding.
One thing Alexandra Kurland often says is that “Everyone will eventually sit down.” This actually comes from horse trainer John Lyons. At clinics, Lyons would tell everyone in the audience to stand up. Then, he’d keep on teaching the clinic. Eventually, everybody in the audience would sit back down. Trainers don’t have to force, push or prod in order to get the behaviors they want. You can set up situations so that the horse will choose on his own to do a certain behavior.
For example, imagine a trainer who has taught a horse to target. She starts by always presenting the target to the horse at the “12 o’ clock” position, right in front of his nose. Once he completely gets the hang of this, she starts asking him to target at 10 o’ clock, a bit off to his left side. If she does enough repetitions of this in a row, the horse is probably eventually going to swing his hips over a step to straighten out his body. After a few repetitions, the trainer could switch to clicking for the movement of the hips.
In this hypothetical example, the trainer set up the horse so that the horse wanted to reorganize himself in a slightly more comfortable position. There’s nothing threatening or scary to the horse about having to turn his head. However, after a few repetitions, he might realize that he’d be more comfortable if he straightened out his body. The horse finds a new, easier way to touch the target and in the process moves his hips, which is the next behavior we want. This involves very low levels of negative reinforcement.
We talked just briefly about an example similar to this at the clinic. But, it left me thinking a lot about negative reinforcement and how we can successfully use negative reinforcement as clicker trainers. There can be a lot of benefit of using small amounts of pressure as information when teaching cues. Many of the exercises at the beginning of Alexandra’s training program use pressure cues to create a language and communication system that will come into play much further down the road.
For example, Alexandra Kurland suggests teaching both stationing on a mat and head lowering using rein pressure during the teaching process and as part of the cue. For stationing or head lowering by themselves, it probably doesn’t matter if you train these behaviors with pressure, targeting, free shaping, etc. However, the way these foundations are taught really comes into play when we start combining these behaviors into larger units and combining them with other piece of the puzzle.
Many traditional horse trainers complain about clicker training because they say you can’t teach flying lead changes or other complex, upper level performance behaviors using a clicker. Many of these people think clicker training means they have to sit around and wait for the behavior to happen. And most average horses are never going to just offer a flying lead change. However, we can use clicker training and pressure as information to build a language of subtle rein, leg, and weight cues. When we have enough of the puzzle pieces filled in, complex performance behaviors will start to fall in place almost seamlessly.
Tabletop Games for Improving You Shaping Skills
Sometimes during training you will end up confusing or frustrating your horse. This happens to all of us, both new trainers and old trainers alike! Clicker training and horse training are both skills that take time, patience and practice to develop. The best trainers I know are ones who have an attitude of continual self improvement. If things don’t go as planned they stay calm, make the best of it, and figure out what to do differently the next time.
There are many ways you can practice training horses without actually using a horse. In the first post of my clicker training clinic notes I wrote a lot about how we used “human horses” to simulate ground work exercises at the clinic. You have probably heard of and maybe even played the Shaping Game, where one person clicker trains another to do some behavior. (If you are not familiar with this game, there is a great explanation of it in Karen Pryor’s wonderful book, Don’t Shoot the Dog.
One variation of the shaping game, which I learned originally from Kay Laurence, involves two people, a collection of small household items and trinkets, a clicker, and “treats.” (The treats can be small objects such as stones or coins.) The two people sit at a table and one trains the other to “do something” with the objects. This can involve shaping different movements with different objects, putting certain behaviors on cue, and even training complex concepts and skills. This is a great way to improve your timing, mechanics, shaping skills, and observation skills.
We spent some time on the second day of the clinic playing some of these tabletop shaping games. Sometimes when clicker training horses (or training any other animal) the trainer thinks she is training one skill, but the animal is actually learning an entirely different behavior. Tabletop shaping games are helpful because we can talk to the learner after the training is over. This is a great way to figure out what the learner was thinking or feeling during the training. Analyzing this information can help you figure out ways to improve your training in the future.
For example, for one shaping problem, I trained one of the other women at the clinic four different cues, two “nouns” and two “verbs.” That is, my final goal was to be able to specify a certain action and then to also be able to specify which object she should use for that action. We had a set of children’s colored markers, so the two objects that I put on cue were for her to select a green marker or an orange marker. The two action behaviors I put on cue were rolling the marker along the table and spinning the marker back and forth on the table.
I was able to fairly quickly train my learner to roll or spin a marker and then put these two different actions on cue. Then, I trained her to select either a green marker or an orange marker and I put selecting each color on cue. Then I tried combining my cues—first giving the cue to select a specific object, and then giving the action cue. This worked perfectly for the orange marker but failed miserably for the green marker. Oops, what was going on? I wasn’t sure, but I went back to practicing each of the four component cues individually. Then, I practiced giving the cue for “orange” with each of the action cues. After a bit of this, I tried again to give the cue for green and then the cue for a specific action. And she got it! From then on, I was able to specify which object and which action.
While discussing afterwards, we figured out what had happened. My learner had been selecting the orange marker by touching it with one finger. However, she had been selecting the green marker by putting her fingers on the table on either side of the object, but not touching the object. (This was a result of the shaping process and her history of reinforcement. She had started by trying to pick up the green marker, which I didn’t want. So I had shaped her to lower her hand down over the green marker, and then put her fingers on the table on either side.)
I thought I had put “pick the orange object” and “pick the green object” on cue. And I had, for orange. However, for green, my learner had thought that she was learning an action for the object, not just an object. The action she thought she was supposed to be doing was to put her fingers on either side of the marker but to NOT touch it. So, later when I ask her to either roll or spin the green marker, to her, I was giving two completely opposite cues. I was telling her to roll or spin the green marker, while, at the same time, I was telling her not to touch the green marker. No wonder she was confused at first!
If an animal has trouble learning a new behavior or concept, it’s not because the animal is unintelligent. During training, our cues might not make sense to the animal because we might be giving confusing or conflicting information. In these situations, we need to return to a previously trained behavior, see if we are missing important component skills, and find a new or better way to communicate with the animal. These types of shaping games are great because they will help teach you how to break behavior down into smaller parts and see what component skills are necessary to build complex behaviors.
Clicker training gives us tools to create happy horses. This clinic gave me a ton to think about in terms of how I can use clicker training to create solid cues and beautifully balanced horses. As well, I was reminded that clicker training gives us ways to make training fun while, ultimately becoming better trainers. I hope you enjoyed this series of posts! You can find links to all three posts from this clinic, as well as notes from other clicker training conferences on my clinics and conferences page.