Jump to content

What Is A Reward (Or A Punishment)?

Posted by vwkoch, in horse handling thoughts 31 January 2012 · 884 views

Most people know that training involves rewarding what you want and punishing what you don’t, so why is everybody not more successful at training? I believe it is because people are not really aware of what constitutes a reward or a punishment. So, I thought it might be interesting to post a blog entry on the subject.

The short answer to the question posed in the title is that a reward is anything the horse wants, and a punishment is anything it fears. It still sounds pretty simple, right? To clarify the problem, let me rephrase the answer just a bit. A reward is anything the horse wants (not just something we THINK the horse wants), and a punishment is anything it fears (not just something we THINK it fears). In other words, training becomes difficult when you THINK you’re using rewards and punishment but you’re really not.

For example, consider the pawing horse that gets yelled at, or even slapped, but it still keeps pawing. Horses are social animals and don’t like being tied up and left alone. They may originally paw out of frustration, but when the pawing gets attention, they have been rewarded, so the pawing continues. WE think the yelling is punishment, but THEY think it is a reward. Do you see the problem here? Even if you slap your horse, the likelihood is that he sees it more as reward than punishment. If you’re slapping him, he’s obviously not alone any more, and he’d rather have “mean” company than no company at all. Even if you ignore him, the pawing helps relieve his frustration, so he’s rewarded no matter what you do --- which is why pawing is so difficult to stop. [Note: If you start by rewarding him for standing quietly for SHORT periods, which you GRADUALLY lengthen, you may avoid the problem altogether.]

Another common mistake is to think that a horse finds a pat or a “good boy” to be rewarding. These signals of approval are HUMAN signals, and they are entirely meaningless to a horse --- unless the horse has been TRAINED to see them as rewards. Whenever you’re with your horse, you’re training it, whether you realize it or not, which is why horses so easily develop bad habits. However, accidental training is not ALWAYS bad. If you usually pat your horse, or tell it “good boy”, right before you give it a REAL reward, then the pat or “good boy” becomes rewarding by association.

Right now, there are probably a lot of people thinking “I’m sure my horse finds a pat to be rewarding, but I’ve never given him a treat after patting him, because I don’t give my horses treats at all.” Again, such thinking is HUMAN thinking. Yes, horses find treats to be rewarding (if it’s the RIGHT treat), but treats are not the only possible reward. Horses are inherently lazy, so if your horse does something good and you let him stop while you pat him, the stopping is the reward that becomes associated with the pat.

Rewards are also called “reinforcement”, and stopping work is called “negative reinforcement”, as opposed to a treat, which is “positive reinforcement.” The terms “negative” and “positive” are not moral judgments --- they are used mathematically. When you stop working, you are SUBTRACTING something (work), and when you give a treat, you are ADDING something (the treat). Horses have historically been trained with negative reinforcement. The point here, though, is that you can reward a horse by taking away something the horse doesn’t like, and under those circumstances, many people may be rewarding their horses without even knowing that they’re doing so. Whatever behavior occurred right before the reward is the behavior you’re encouraging, so again, it is no wonder that horses so easily develop bad habits.

One reason why we usually use negative reinforcement to train horses is because they are a prey species, as opposed to dogs, which are a predator species. For predators, food is difficult to obtain and comes in spurts, so food is very important to them. Because food is important, treats make a good reward --- predators are almost always willing to work for treats.

For prey, food is usually readily available and fairly easy to obtain. Grass is normally not a scarce resource. Prey species still find food important, but it is not as much of a priority for them as it is for predators. For prey species, unless they are starving, safety is a more important consideration than eating. Domestic horses are pretty safe overall, but it is possible to threaten that safety (by waving a whip at them, for example) and then reward them by removing the threat. That kind of reward is negative reinforcement, and the prey vs. predator background is why it works so well with horses.

Certainly, horses will work well for treats, or other positive reinforcement, and dogs can be trained with negative reinforcement, as well as with treats, but it is useful to understand some of the reasons why our historical training methods are what they are. We train horses with negative reinforcement because it is easy and it works. However, it also increases the possibility of the introduction of abuse.

Consider the use of typical round pen training. The horse is made to run around the pen until it does what the trainer wants, which is to approach him. The horse learns that, even though the trainer is the threat, the safe place to be is right next to him. So, the horse wants to be next to the trainer not because it has developed any liking or trust of the trainer, but simply because it can avoid his threats by being near him. I just can’t see that approach as being the basis of a good relationship.

In addition, if the trainer is a poor one and does not time his (negative) rewards properly, the horse will have trouble learning what is expected of it. When the training isn’t working, the poor trainer becomes frustrated and increases the threat, thinking that the solution is to make the horse work harder so it will have a greater desire to rest. However, the problem is not that the horse doesn’t want to rest. The problem is that it doesn’t understand what to do to be ALLOWED to rest. Making it work harder won’t solve THAT problem, and running a horse around a round pen until it is exhausted is not training. It is abuse --- intended or not.

The benefit of positive reinforcement training is that, if a poor trainer doesn’t time the rewards properly, he will just be giving the horse more treats, not more punishment. His poor methodology won’t lead him to abuse the horse before he recognizes that his approach isn’t working and he needs the help of a better trainer. However, changing from negative to positive reinforcement methods is a BIG change, and I don’t believe a change of that magnitude will happen any time soon with regard to horse training. Therefore, trainers must be aware of the possibility that the use of negatives can degenerate into abuse, so that they can make sure that their methods do not become abusive.

Although I think positive reinforcement is better, negative reinforcement is not inherently BAD, and I do use it myself. However, I use the mildest negatives possible, and I couple them with positive reinforcement. I use mild threats to make my horse run around a round pen for exercise, then I give her a treat when she’s done. My horse considers me her safe place not because she has to come to me to avoid my threats but because she knows she can come to me for protection from scary things in the environment. She also likes me because I am a source of treats. We have become friends, and I don’t think that kind of relationship can develop with a trainer who uses ONLY negative reinforcement. If you’re always harassing your horse, you might manage to make your PROXIMITY into a safe place, but you will never be seen as the SOURCE of safety and a person to trust.

Even expert trainers make mistakes, especially when accidental training is happening, but experts can look at a bad habit, consider the past, and recognize where they went wrong. They may not identify subtle rewards and punishments during the accidental training, but they will be able to spot them after the fact, and correct whatever they did to cause the problem. It’s not as easy for the average person to figure out what happened and how to fix it, but it helps to know that bad habits ALWAYS occur because they are rewarded, so you need to look for the reward in order to be able to take it away. It also helps to know that you can’t train a horse by providing it with a reward that only YOU think is a reward. It’s only really a reward if the HORSE thinks it’s a reward. So, when you’re training, remember to try to think like a horse, not like a human!

Thank you for clarifying this subject. The article was well written and lots of thought was involved. As I was reading the posted article, I thought it could have been written by Julie Goodnight or Clint Anderson.
  • Report

Thank you for clarifying this subject. The article was well written and lots of thought was involved. As I was reading the posted article, I thought it could have been written by Julie Goodnight or Clint Anderson.

  • Report
Not meaning to be critical, but you seem to have a lot of things half right. Those who train lots of horses have the knowledge to use feel, timing and balance. Not my words, but the words of Tom Dorrance. using the round pen is something that involves skill to actually teach a horse, by using pressure and the release of pressure. It is not about making the horse run to Avoid a threat, it is about the horse responding to pressure and the release of pressure. I have watched lots of trainers do exactly as you described and get the horse tired to achieve their goal. That is not really good round pen training, the idea is to work their mind while giving them the freedom to move if they are not ready for it. It is not about getting them tired, but giving them choices. Horses need to be an active partner in the horse/man relationship. To accomplish that they need to know from the start that they have choices to make in this relationship. They can choose to respond to the release of pressure and stop or step closer, or they can choose to move. I will teach them to turn in towards me by the release of pressure. You are correct that you are training the horse, positively or negatively at all times, if you are meaning to or not. The thoughts of rest being the reward are very important, I don't care how much my students talk to their horses or call them poopsie, their horse does not care. The important part is that they reward the work with rest and make the right thing easy and the wrong thing difficult.

Less of a disagreement and more of a clarification if I may.
  • Report
People who talk about "pressure" are talking about threats, whether they realize it or not. "Pressure and release of pressure" are just another way of saying threat and removal of the threat (i.e., negative reinforcement). When you say "the idea is to work their mind while giving them the freedom to move if they are not ready for it", you are essentially saying "threaten them and allow them to move to try to escape the threat if they don't know how else they can escape." A good trainer then removes the threat when the horse makes the slightest move toward the desired behavior, so the horse learns that it can escape the threat by doing something other than running, but a bad trainer never notices those slight moves and can run a horse to exhaustion. Horses do have choices, and they can choose to run or to stand when they feel threatened, but if they REALLY had a choice, I suspect they would choose not to feel threatened at all.
  • Report

October 2016

23242526 27 2829

Recent Entries

My Picture

0 user(s) viewing

0 members, 0 guests, 0 anonymous users