I have often spoken with horse owners about Operant conditioning training and the different forms of teaching we can employ. During these discussions, I have sometimes struggled with making clear the arguably subtle distinctions between the methods employed in classical training and those more often used today. Yesterday I had such a conversation and as I thought back, I realized that even though I was speaking with someone who understands, much better than I do, the technical definitions used in psychology to delineate different forms of learning, I still found myself stumbling as I tried to explain how my take on ‘classical training’ differs from a lot of other forms of training. What I hope to do in this article is to better articulate this.
For the sake of this discussion, I must clarify that what I want to talk about is the how, of teaching the horse rather than what. So let me sum up the what and get it out of the way first, so we can focus on the how.
In ‘classical training’ we seek to balance the horse. Too vague? Okay, what I mean by ‘balance’ is physical, mental, and emotional balance. Physical balance is where we wish the horse to carry the rider with the center of balance directly under the rider and supported with even supple strength from the left and right sides of the horse; through the aids, we are then able to affect the physical balance to direct and guide the horse. For mental balance we want the horse mentally engaged without being willful; alert to the aids and rider’s will, without becoming a robot; and responsive but not trying to guess what is coming next. Emotionally balanced is that they are calm without being dull; alert without being excited; trustingly obedient and confident at the same time.
“Well, of course, we all want this from our horses. How does this differ from the goals of ‘non-classical training’?” Excellent question. First, let me say most riders don’t understand the importance of the horse being physically balanced and when they do, few understand how to accomplish this balance. Mental balance is often not made a priority because the mind of the horse is understood by even fewer riders than the physical form. Emotional balance is seldom achieved because it requires first that the emotional balance of the rider be firmly in place and little factors like fear, ego, enthusiasm, and impatience tend to get in the way of that. All this notwithstanding, the real difference is how we go about achieving what we understand as the goals of horsemanship.
The methods used to help the horse achieve physical, mental, and emotional balance we seek fall into a couple of broad categories of learning; Operant Conditioning and Respondent Conditioning.
In Operant conditioning behavior is modified by consequences. This form of learning takes aim at voluntary behavior, affecting it by affecting the environment of the horse and is maintained by consequences.
Respondent conditioning, (ironically, also called Classical Conditioning), deals with the conditioning of reflexive behavior through antecedent conditions and is not maintained by consequences.
“I don’t have a degree in education or psychology so what does all that mean?”
To simplify, Operant conditioning is: something happens, the horse responds or doesn’t respond, there is a consequence. For example, you pick up your right rein and increase contact on the bit, the horse finds this less comfortable and tries to pull to the left, but the contact remains, when it moves its head to the right, you lower your hand and the contact returns to something comfortable. Repeated consistently, stimulus, voluntary reaction, removal of stimulus, the horse learns to move its head to the right with the right rein is lifted even before discomfort. If however, you don’t always release the contact when the horse turns its head to the right, it will begin to seek out another reaction to the pressure to solicit a release, and the training is broken. This method of training is more often used as it deals with the voluntary and therefore physical actions of the horse and in turn, through how we respond to these reactions, the mental state of the horse.
Respondent conditioning works on the reflexive response to a stimulus. For example, we all touch or rub our horses to calm them, much as their mother did when they were foals. Because of how the horse has evolved, there are points on its body that cause the release of endorphins when rubbed. If while working the horse ‘in hand’, we gently rub at the withers and speak softly as we bring the horse quietly to a halt and do this over and over and over again, that animal will begin to reflexively associate speaking softly with comfort and this will cause similar feelings to standing quietly and being rubbed on the withers. Once the condition is in place, you don’t have to speak softly every time you stop for the reaction to still be there on the times you do. As this method of training deals with reflexive responses, it can be employed to affect the emotional state of the horse. Tapping into the ‘hard-wired’ aspects of the horse’s nature, we can condition emotional reactions to engender trust, respect, and confidence, but we can also create fear and self-preservation responses without intending to.
It is important that I go into more detail now about the contexts of operant conditioning. I am going to be using the terms positive and negative but not in the way they are most commonly used. Positive does not mean good and negative bad, but rather we are simply referring to addition or subtraction.
Positive Reinforcement refers to giving the horse something it wants when it does something you ask, or in other words, a reward-based method of training. The training style called “Clicker Training” uses this method as its primary focus, though it is combined with respondent conditioning to associate a reward with the sound of a clicker, and the clicker is then used, in place of a treat, when the horse does what it is being asked. For example, they touch the horse’s front left foot with a stick and say “Up”, if the horse lifts its foot they click the clicker. It is the same when a horse is taught to ‘bow’ by offering it a treat and then the treat is moved in such a way that the horse, in trying to get at it, attends the position the trainer is looking for before the treat is relinquished.
Negative Reinforcement is the most commonly misunderstood term in the training world. This context refers not to doing something ‘bad’ to the horse when they don’t do what they should. Rather, it refers to the removal of stimuli when the correct reaction is given. The trainer provides stimuli to cause the horse to react and removes it the moment the horse responds the way the trainer wants. The example of operant condition I gave above refers to precisely this method. Squeezing the legs, the horse to move forward, relaxing the legs is another. Holding the reins firmly in place to cause the horse to stop, then releasing the firm contact as they stop, is yet other example of negative reinforcement. This is the method most commonly used in classical training.
Positive Punishment is not the oxymoron it appears to be and is very common in the horse training world. It refers to the addition of a consequence for an incorrect response to stimuli. The trainer pulls back on the reins to stop the horse, but the horse does not stop as wanted so the rider gives the reins a ‘pop’ to punish the horse for not reacting as requested. In my view, there are times when this method must be applied, especially where dangerous behavior is the issue, as in smacking a horse for nipping or even striking a horse with a crop when it attempts to walk over you. That being said, in classical training, we hold this as the last resort.
Negative Punishment is the fourth context of operant conditioning and calls for the removal of something desirable when the wrong response is given; taking away a child’s toy when they demonstrate an undesirable behavior for instance. This method does not work with horse training, because it requires a rudimentary abstract awareness of cause and effect that is beyond the horse. Though I have seen people try it; “No, you don’t get a treat today because you refused to go through the water.” I feel this to be completely useless.
Classical training opts for negative reinforcement as the primary operant conditioning method it employs to achieve the physical reactions we seek, with a positive mental state, while doing as little as possible to upset the emotional balance. We will combine this with positive reinforcement to speed the process along and to help form the proper bond between horse and rider. The classical quote “Ask often, be content with little, and reward lavishly.” speaks to that exact balance. That is not to say we never use positive punishment in classical training. As I have said, where safety is a concern, it is clearly warranted, but there are other times when I feel it is the correct method, though most trainers are unlikely to use the word punishment. My particular euphemism is ‘focusing the horse’. When I feel a horse has become distracted, preoccupied outside the ride, or willful, I will give a light tug on the rein or touch with the whip to ‘refocus’ the horse. What is important to note here is, that it is the loss of attention I am punishing not the incorrect reaction to the aid. Such ‘focusing’ moments are usually brief and the return to the preferable negative reinforcement is immediate. I would say that the most important thing to keep in mind about positive punishment is that it can only be used as a method to negate an action and should never be used when a horse might simply not understand what is being asked of them. In other words, just because they got the wrong answer is no cause for punishment. It is only when with they are doing is a behavior that must be dissuaded that we punish and then we must only punish the act, not the horse.
Many trainers use positive punishment as the primary method without ever really giving it a great deal of thought. Yes, they can attain the physical response that is essentially similar to the classical trainer, but in the process, damage the mental and emotional balance we seek, causing the horse to become tense, shut down, or become untrusting. “Ask first, then demand.” or “Offer a good deal first, then give them the ‘not so good’ deal” These quotes offer us useful insight into this method. The trainer requests the horse bend to the left with a very light guiding hand. If the horse does not bend, the trainer ‘pulls’ the horse’s head over with force to get it to comply, then releases it. The process is repeated until the horse realizes that the more painful pull is going to be the punishment for not responding to the softer one. While this method can be of wildly varying harshness, with some trainers causing wounds on a horse’s mouth if it fails to stop fast enough, for instance, the context of the basic is the same; ‘Ask, demand, release, repeat.’ The problem as I see it, is that this method is based essentially on fear. Training with fear certainly can work in that it teaches certain types of horse to react to the light aid, but that reaction does not come from trust, and without trust, you cannot have légèreté and without that nothing is classical.
At the other end of the spectrum, those who use positive reinforcement to excess run the risk of losing the respect of the horse and finding they don’t have the ability to influence the horse’s actions when facing never-before encountered challenges.
It is important I think to remember that the choice of operant conditioning method and how it is employed can have an intentional or unintentional respondent conditioning effect. In the process of gaining the physical responses to the aids we seek, we can also create reflexive reactions for the better or worse. A rider can ask his green horse to stop by pulling back on the reins and saying “whoa” but when the horse fails to understand or keeps moving due to miscues with the legs or balance issues, the rider can get angry and pull very hard on the reins and yells “I SAID WHOA”. The horse, to escape the pain of the bit being pulled against the tongue, lifts its head way above the bit and hollows its back, but keeps on moving because now it is scared and in pain. The rider gets frustrated and insists the horse is going to get this right if it kills him. By repeating this same basic dance of failure, time and time again, eventually, the rider finds that the horse panics and raises its head every time the rider says “woah”, even if the reins are laying slack across its neck. This is an extreme example of the unintended effect and happens to be one I have firsthand experience with.
With the proper combination of operant methods, applied thoughtfully to achieve the wanted physical reaction, a desirable secondary effect can be achieved on the horse’s mind and through respondent conditioning, a similarly desirable tertiary effect on its emotional response. These subsequent effects combine to create a synergy in training, each one helping the others, in an upward spiral.
Inversely, the wrong choice of operant method can create undesirable secondary and tertiary effects causing the spiral to go downward, and as this obvious mental image would suggest, once it has reached a certain point there is a little chance of stopping it.
The earlier one is aware of how one training method is affecting the horse as a whole, the greater the chance of using it to influence the training for the better. It is the making of this awareness a cornerstone of training that makes it classical or not.