Wednesday, August 1, 2018

Saying "Don't!" With Positive Reinforcement

Last week I wrote a post on my Dog Chapter blog about Saying No With Positive Reinforcement. This week I'll address a similar question I hear, which is how to prevent a horse (or other animal) from doing something. A common example is:

"I need to teach my horse never to go over the top of me"

Not only is this about wanting "obedience", but in this situation, there is definitely a fear factor.  Having a horse run over you is a very real and dangerous possibility when dealing with large fearful animals. People want to know how to prevent that from happening. And if it looks like it might happen, then what?

First we must be sure we are listening. By listening, I mean observing the animal's emotional signals so that we can respond in a way that keeps everyone safe and happy. In the case of horses, we must know what those emotional signals look like, and we must also know our individuals for more subtleties. Once I see that the horse may be getting worried, my positive reinforcement response is the same as saying "no". I teach what I want instead. In this instance, it must be practiced, practiced, practiced so that I can access that response when tensions are high. 

And remember, even humans who are panicked will run over other humans with disastrous results.  There are no guarantees in this life and if you choose to spend some of your days with horses, you must accept that it's potentially dangerous.

As a devotee of Alexandra Kurland, I look to five of her Foundation Lessons to assess my horses' responsiveness. These are behaviors that my horses know solidly and have practiced for years so their reinforcement history is strong. Each of the lessons can be used a safety net, IF they have a reliable response. 

I remember well the clinic I attended with Alex when she helped me begin to take the "make it happen" out of my cues. Alex's clinics frequently, if not always, use "human horses" in practice.  Having a human tell you how it feels when you grab the lead rope is very enlightening. I had no idea how forceful my requests were and I needed to work on changing my demands to requests if I wanted a quiet response from my horses. 

"Make it happen" demands only work when you are scarier than other things in the environment. If someone tells me to sit still when there is a large rock rolling down a hill toward me, it doesn't matter how forcefully they say it, I am going to get out of the way of that rock! Unless, I trust that person.

Trust is a messy word when it comes to horses.  I see so many people saying that horses trust their handlers: horses will lie down, or go into water, or jump big fences. In some instances it may be true. In other instances it may be because the horse has been taught that he doesn't have a choice. The fear of what might happen if they don't obey is actually stronger than the fear of doing what is asked.  So in the case of the rock coming down the hill at me, I need to trust that the person telling me to sit still knows the landscape well enough to know that the rock will bounce and go around me instead of hitting me! 

The more experiences I have trusting this person regarding bouncing rocks, the more my trust in them will build. I may sit still in otherwise ridiculous situations (almost as ridiculous as this example) because that person has shown me that trusting them keeps me safe. We need to give our horses similar histories of trusting us in low risk situations before we ask them to trust us in higher risk situations.

This morning I took my Percy horse out to an area he deems somewhat scary. He is not imagining things- as peaceful as it looks in this photo below, we have seen quite a variety of wildlife come out of those woods, from innocuous deer to coyotes and bears. I know his hearing and other senses are sharper than mine- I trust that when he alerts to something, something is really there.  This is because history has shown when I see him standing in his paddock on alert, I can run to the other side of the house and look out the window and see interesting wildlife. Trust goes both ways. 



This morning I was working on building trust that I won't ask him to do anything he thinks is dangerous. When we first got out here, I walked up to a mat which I had previously placed in the grass. My question was, "can you step right onto this mat in this environment?" 


He could. I noticed a slight lowering of criteria in the quality however. That is an important  observation.  In addition to the height of his head and his intense focus on the woods (which you can't see), his feet did not land squarely in the middle of the mat, nor did they stay square with each other. This could have been due to two different things (or something else altogether). Either he wasn't 100% focused on his behavior to the criteria he can exhibit in a more comfortable environment, or it could have been because of the placement of the mat- on a hill, on deep grass. It might have been a little unstable or he might have been a little unsure.  While I wasn't sure of the reason, it was important that I notice this lack of meeting criteria and proceed carefully, rather than just forging ahead. 

I proceeded through the other foundation lessons, cueing, observing responses and making mental notes. Could he back up at the slightest suggestion of a cue? 



Yes. 

Could he target my hand?



Yes. 

Could he stand with his feet still next to me for a count of 10? 


Yes.
Could he drop his nose to the ground and leave it there?



On the second try. He put his head down on the first request, took a bite of grass and picked it up again.  So here is another set of questions for me.  Did he pick it up because he was uneasy or because that bite was not to his liking (my rules are that you can eat when I ask you to put your head down as long as you don't move your feet). When I re-cued, I only asked for a duration of five seconds. 

So what all this tells me is that I have my horse in an environment that he might not be completely comfortable, but one in which he trusts me enough to respond to all my cues, rather than an "every man for himself" situation. Importantly, he responds willingly, immediately and for the most part meeting criteria. 

It's important to keep in mind that the behaviors he has done are behaviors that are very, very easy for him. I am not asking him to perform advanced trigonometry, but recite his A,B,C's (which he has known for years). Next I am going to ask him for something a little more difficult, especially in this environment, and in the way I ask him. 

One thing I know about this horse, and probably many others, is that he'd much rather be looking toward the potentially dangerous area than have his back to it.  Horses are horizon scanners. They want to know when a potentially dangerous predator comes over that horizon so they can be well on their way in the other direction. Asking Percy to turn his back on the scary area invites a scoot forward (or two or three) as he hears things behind him and can't see to determine the risk. 

In previous days, he has been fine as long as he is facing the woods, but each time he has scooted forward as we walk back toward the barn. Here is a situation where I want a "please do not go over the top of me as you are scooting forward".  If we are walking away, I know the scoot will just take him forward a step or two.  But it's that turning around where there is potential for danger. Horses are good at turning on a dime.  If he is to the right of me and turns on a dime toward me at a high rate of speed, that takes him over the top of me. So what do I want instead?  I want him to make a small arc around me as he turns. And I want him to stay balanced over his own feet, rather than leaning in (which if the footing is at all slippery could cause him to slip and end up on top of me). In comfortable environments, he moves around me just like that.  But in a situation where I am asking him to turn his back on the woods, I want to be clear in my cues that I want him to go around me. 

Fortunately, using another of Alexandra Kurland's lessons, I have taught him how to turn exactly like that. If I hadn't previously gone through a training process to train this, it would not be fair to ask him to do it. In this short clip, you can see the first time I ask him to pick his head out of the grass, wrap his neck slightly around me, and step toward the woods with his shoulders, taking those shoulders away from me and preparing him to turn in that arc around me. I only ask for a step or two this first request. I am not "making" him get out of my space.  I am asking him if he can do it. 


The answer is yes.  

So this is my process.  Are all these behaviors solid in a comfortable environment?  Yes. Are they solid in an environment where he is less comfortable?  That's what we are working on. His responses today are a result of training I have done on previous days.  Today's training prepares us for future days. Setting him up in a slightly worrisome environment gave us both needed practice.  I need to be able to reach for cues quickly and he needs to be able to respond to them. 

The next time we come out here, I will see whether asking him these questions today made the responses stronger or weaker.