When It Comes To Training Horses, 'You Get The Behavior You Reinforce, Not The One You Want' - Horse Racing News | Paulick Report

When It Comes To Training Horses, ‘You Get The Behavior You Reinforce, Not The One You Want’

Much has been written through the centuries about the process of training horses, much of it specific to the type of work a horse is meant to do. In a recent webinar hosted by the British organization World Horse Welfare however, experts reminded horse owners that it's critical to take into account the way horses learn and process information when setting up a training program for them, regardless of the job they're intended to do.

Dr. Gemma Pearson, veterinarian and equine behaviorist, said that horses do not learn the way we do. As a species, they have what Pearson called “limited mental capacity” which isn't to say they aren't intelligent, simply that they learn best when complex tasks or situations are broken down into very simple steps where it's clear what they're being asked. Pearson used complex dressage movements as an example. Many of them start with a horse learning two different cues from a rider's leg — speed up, or lengthen stride. It helps horses to feel the rider use different part of the leg for each request, so it's clear what's being asked. The same is true for rein cues, which can be broken down into different but related questions. As a horse's training advances, a rider can combine these clear, well-learned instructions for more complicated results.

“If we teach each response independently, and make it very obvious to the horse what it is that we want, we can then start to put different aspects of that together,” she said.

Horses benefit from clarity, where the correct answer is easy for them to get, and that often means breaking a task or problem down into small pieces, remaining patient, and rewarding the horse immediately for a correct response or even an attempt at a correct response.

From there, trainers must reinforce desired behavior properly, but it's important to think about what reinforcement actually is. Learning theory incorporates several types of reinforcement, but the two most effective with horses are positive and negative reinforcement. Positive reinforcement is the addition of something pleasant, like a treat. Negative reinforcement is the removal of something undesired as a reward, like the removal of pressure from a rider's leg after the horse begins to move forward.

Pearson pointed out that what we consider “positive” may not always be positive to the horse. Food rewards are shown to work quite well, and scratches on a typically itchy area like the withers mimic the bonding grooming that horses practice with each other. Patting a horse however, is probably slightly confusing to the horse, as it doesn't resemble any kind of communication between horses and if anything is more like a person tapping them as a signal to move. Pearson also expressed doubt that vocal praise is necessarily intuitive for horses either, as horses may pick up on tone of voice but not the specific meaning of a phrase. Negative reinforcement could include the removal of pressure from a leg or a hand, but it could also include a short break in a training session.

Horses are also very susceptible to classical conditioning, the well-known premise illustrated by Pavlov's dog. The dogs in Russian Physiologist Ivan Pavlov's experiment learned that the ringing of a bell meant food was coming, an association they drew so clearly that they began salivating when he rang the bell even when there was no food on hand. Pearson pointed out that horses can also learn to anticipate very well, and that ability comes out in all kinds of ways. Under saddle, it means a horse can gradually become more sensitive to shifts in a rider's position as they prepare to ask for a movement or transition, and will soon react to those shifts without needing the actual cue.

Of course, this level of sensitivity has its drawbacks.

“You get the behavior you reinforce, not the behavior you want,” Pearson pointed out.

Sometimes, it's not immediately clear to the person involved that they are reinforcing an undesired behavior. Pearson used the example of a horse that stretches its head up to avoid taking medication from an oral dosing syringe. The horse has created his own negative reinforcement there — raise head, syringe disappears. The training goal in that situation shouldn't be to prevent the head raise, but to make the horse want to keep the head low and tolerate the presence of the syringe. Pearson suggested a combination of positive and negative reinforcement there, by providing a treat when the horse kept his head low and remained calm, and also by placing the syringe near the horse's mouth and removing it when the horse remained still and calm.

Pearson has consulted on a number of cases of problematic behavior or training challenges through the years and finds that they usually come down to a few core problems. The most common one is undiagnosed pain, which Pearson estimated impacted 80% to 90% of the cases she has seen, and is easily missed if it's not pain resulting in a clear, asymmetrical limp. Other factors can include situational stress or mental stress in a horse's living situation that makes it difficult for the horse to focus on the training session.

“There are no bad horses,” Pearson said. “There are lots of horses where pain is causing problems, there are lots in not-great environments, and there are certainly lots of badly-trained horses.”

Watch a full replay of the webinar below.

Paulick Report Icon

Receive daily headlines, breaking news alerts, promotions, and much more!

Ask Ray?

Have a question for Ray that you'd like answered? Have a news tip? Ray will go to his network of sources in pursuit of an answer.

Ask Ray