LEARNING THEORY FOR HORSE TRAINERS
Few of us buy our equine partners fully trained. Most of us choose a horse that we think has the temperament and ability to the job we are interested in and train them ourselves. But most horse riders don't think of themselves as animal trainers. There is a vast body of knowledge from 100 years of academic ethology (the study of the behaviour of animals) and psychology about how animals learn. Our ignorance of this science is in great contrast to dolphin, circus and dog trainers. Professional dog trainers teach the rudiments of learning theory to dog owners, even children, so that they can continue to be effective trainers of their pet. This approach is more than often absent in rider coaching. Our horses are the poorer, with confused badly trained horses ruined, and we are the poorer as we achieve less than we could and take longer to do it. Internationally recognised Australian trainer and academic Andrew Maclean makes this point strongly in his book. Good training is an animal welfare issue. American showjumping guru George Morris makes a similar point when he said “if you care for your horse train it well”. A well trained horse will have a good life.
It is easy to be put off by the jargon in the learning theory texts and websites, but a little attention to this is worthwhile. The jargon words will be highlighted in the following discussion.
A great deal of horse training is based on application of pressure (the aid e.g. kicking to indicate go forwards) and release of pressure when the horse responds in the way we want him to e.g. goes forward, or sideways etc. PRESSURE-RELEASE training is also known as OPERANT CONDITIONING. The horse works out what we want by trial and error e.g. at first he may stand still, or move sideways or backward, or even buck or rear. But we keep kicking, until he moves forward. We keep the pressure on, kicking. When he moves forward we take the pressure off, thereby rewarding him. It is the release of pressure which teaches him, and repetition which trains him to always respond to the aid.
It is critically important to understand that the reward, pressure release, teaches the horse. If a timid rider backs takes the pressure off, stops kicking, because the horse is naughty, the rider teaches the horse to be naughty. Similarly if the horse moves forward and the pressure is not released the horse does not learn to respond to the leg. If for instance you are teaching lateral work, say leg yielding and the horse does a step or two correctly, and you, the greedy trainer keeps the pressure on to get some more steps, you lose the opportunity to reward and teach him that what he did was correct. So then he may try something else to get rid of the pressure, and it may not be the steps you wanted. He may shake his head, pull the reins out of your hands and while you are attending to this you take the leg aid off, and bingo, you have started to teach him to reef to a lateral leg. We must be quick to reward (release pressure) if we want to be effective trainers. We must not be greedy, but should bear in mind the concept of SHAPING. This means starting with a small response, e.g. one sideways step, then slowly refining it until it is several steps, in rhythm, etc by repetition and reward. The more you reward the more chance he has to learn.
When shaping behaviour the trainer must be careful to
1. Shape one behaviour or quality at a time
2. Gradually reward better attempts at the behaviour, e.g. at first any forwards movement should be rewarded, and later when the response is established only an immediate, straight, rhythmical response may be rewarded.
3. If you get to a point that the horse is not achieving rewards go back to a point where the horse is confident and can be rewarded. Remember that if he is not being rewarded he is not learning.
If he experiences frequent rewards with the release of pressure he builds his confidence in you and in the work, and I believe he enjoys his work and can even develop a liking for his work and a desire to please you.
Now you know from my last article that I don't really think that a horse thinks “this will please her”, because for a start he isn't capable of the required empathy, i.e. he hasn't the capacity to think about your mind state of ‘being pleased'. However, if you are consistent he will know that you will release the pressure, so he becomes more willing. He joins your team, and the aids can become lighter, he gets better at the trial and error process of learning what will release pressure. Andreas Helgsrtand has written about making your horse a team player, and this is what he was talking about.
When you go to a show watch the warm-up arenas. I know in my discipline you will see horses being ridden around and around, the pressure is on. More exercises, more movements, more collection. You will often watch for quite a while before you see the pressure come off. Then go and watch the top horses and trainers. You are much more likely to see breaks with the pressure off. The horse does something good and the trainer rewards. The horse is allowed to walk briefly. The better the horse goes the more often the trainer allows him to walk. As a first season Grand Prix dressage rider I have noticed that when I am warming up for the grand prix many of the other riders also stop and walk often, compared with the warm up rings at lower levels where the less knowledgeable riders grind around and around, so busy concentrating on improving the work that they forget to let the horse know what they want, by rewarding him. The rider is satisfied with the half pass, so straight into the changes without letting him know that he was pleased. So this horse may see little point in trying hard, because it doesn't reduce the pressure any way. The work may become mechanical and lack lustre, or more pressure may need to be applied to get a response, or the horse may become unwilling or disobedient and uncooperative. If you want harmony and lightness, expressiveness, in your work you must remove pressure when the horse tries, and shape the response.
You can train him to respond to a lighter aid in the same way. Apply a light aid. If he doesn't respond apply a strong aid. Release pressure as soon as he responds and repeat the light aid. If he responds well to the light aid you can JACKPOT, the jargon for a big reward. For example, if I am training a horse to do something new and he makes a really good effort I might get off, even before he stops. The reward must be as soon as possible after he attempts the right thing. The first time he does a flying change for example, give him a big pat, say good boy and get off. He mustn't think the reward is for halting after the change, he must still have the change in his mind when you release the pressure.
Now if you are concentrating you will notice that saying “good boy”, and patting him is not PRESSURE RELEASE. Correct. The jargon for those training techniques is a CONDITIONED REINFORCER. A horse learns by association that patting him on the neck or saying good boy means that the pressure will be released. The pat on the neck or saying good boy doesn't mean anything to him until he has learned by repetition that the pressure will be released. Then the pat, or “good boy” can reinforce the training. Pzazz believed for a long time that “good boy” meant stop, and would try to stop as soon as I said it. I had to re-explain this to him, so he didn't stop in the middle of a movement that I want to encourage. A REINFORCER is something that increases the behaviour, a POSITIVE REINFORCER is the addition of something like a reward, food, pat etc, whereas NEGATIVE REINFORCER is the removal of something (yes, like removal of pressure). Note here that negative reinforcement is not punishment. CLICKER TRAINING is the use of a clicker immediately after the horse does something you want, followed by a food reward. The horse learns by the association of the clicker with the food reward that he did what you wanted. A child's party clicker is fine or buy one from a pet shop in the dogs section. This works fantastically well with horses that are motivated by food. I have a Shetland pony who will pick a dressage whip off the ground in his teeth, do Spanish walk, and do a variation on piaffe, all of which he learned quickly with clicker training. He loves this work because he loves the positive reinforcer, the food that follows the clicker. Clicker training has huge potential to speed up training in the horse world; you just have to take the time to establish the association of the click and the food reward. This is easily done by using a target, like a bucket lid. When the horse touches the bucket lid with his nose click and reward with carrots or pellets (I don't like to use sugar). It is ok to help him by moving the lid really close to him to start. When he understands he will reach around or even walk to the target. Once he understands that the clicker will be followed by food you are set to teach him tricks, or to pick up his feet or put his head down to be bridled, or piaffe, whatever you like. Just follow the rules of shaping, and only train one thing at a time. For example, to teach Blake to pick up the whip we first used the whip as a target, rewarding him for touching it with his lips. We then rewarding him for mouthing it, and by trial and error he worked out that he would get the reward only for picking it up. The first time he did that he got the JACKPOT, a big handful of pellets.
So CLICKER TRAINING is a form of OPERANT CONDITIONING which uses the clicker as a POSITIVE REINFORCER.
CLASSICAL CONDITIONING is where a neutral stimulus (e.g. a bell) becomes associated with an involuntary response or reflex so that the stimulus produces the response. The neutral stimulus must be presented before the response you wish to promote. The best known example in the scientific literature is ‘Pavlov's dogs'. The scientist Pavlov rang a bell just before feeding his dogs and after enough repetitions the dogs salivated when the bell was rung, even when the food was delayed. Classical conditioning is how we train horses to respond to aids that are not removal of pressure, eg seat aids.
HABITUATION is simply the process of getting used to something, ie making habits. In the absence of real training this is how horses learn much of what we expect of them. They get used to rugs and saddles and being tied up. A lot of very ineffective training works because we have been breeding horses and selecting for temperament for a long time, and good horses will get used to a lot of things.
We know that horses do not learn by reasoning, or logic, or watching other horses. Human beings can do all of that but horses do not.
Let's take a simple and common example of something we all want our horses to do. Loading on to the float. You can use HABITUATION. We do this with our foals at weaning. The float is securely placed in the yard and the feed is put in the float. The foal is free to put himself on and off and just gets used to it. We don't teach him. But later, although he is familiar with it he may or may not wish to load.
OPERANT CONDITIONING. First teach the horse to go forward from the tapping whip by tapping repeatedly on the stifle or hind leg, and rewarding forward steps by immediately stopping the tapping. When he completely understands this ask him to walk forward onto the float with the same signal, tap, tap, stop as soon as he steps forwards. This asks for PROOF that he will obey the trained signal in a novel setting. Do not take the pressure off until he steps forwards. This is the technique promoted by genius Australian horse trainer Tom Roberts.
If he runs back or rears maybe use a stallion chain, or rearing bit, to prevent those disobediences. It is important in training that you never let him practise being disobedient. But keep tapping until he steps forwards, and stop tapping immediately he moves. If he loads and then runs backwards, never mind, just apply the pressure, tap tap, and do it again. Reward the attempts. SHAPE the behaviour by rewarding small steps forward at first, but once he understands expect a big step, or a few steps before the reward. If he understands the clicker you can use CLICKER TRAINING, tap tap, click and feed to reward the forward steps. Then when he is confident you may wish to use CLASSICAL CONDITIONING and say “walk on” before you tap. Soon he will walk on, in front of you if you wish, on the verbal command. Just remember that CLASSICAL CONDITIONING is not as strong as OPERANT CONDITIONING and from time to time you may need to use the whip to tap to keep him honest.
When you are using the whip remember you are not punishing him for not going forward, rather signalling persistently. You do not want to frighten him to go on, just persist. You may wish to increase the stimulus intensity (use the whip a little harder), but you do not wish to hurt him. Just out-stubborn him.
As I tell my pony club students at the water jump “you are a fine example of the most highly evolved and intelligent species in the world, possibly the universe. You can out-persist a horse. Stay cool”. Remember that the fastest way to train a horse is to make haste slowly and train him properly.
And remember to have fun.
If you want to learn more: THE TRUTH ABOUT HORSES, Andrew McLean, Viking 2003
www.clickwithhorses.com from Australian paralympian Georgia
Technical articles