Learning theory is a scientific principle of how learning takes place, based on the work of behaviourist BF Skinner. Learning is a result of either reinforcing the desired behaviour or punishing an undesired behaviour.
It isn’t just a training theory – all learning is encompassed by this principle. A wild horse who finds water to drink when they’re thirsty is reinforced to repeat the actions that led to finding the water (positive reinforcement). One who is scared by an encounter with a predator learns to avoid similar situations to keep themselves safe (positive punishment).
Using this framework allows us to consider what the impact of our training methods is on our horse, and can help us become more thoughtful and effective communicators.
Learning happens through reinforcement (something the learner finds appealing) and punishment (something the learner finds unpleasant).
Both reinforcement and punishment can be positive or negative. In this context, positive and negative don’t mean good and bad, but instead think of them mathematically: positive means adding something, and negative means taking something away.
Positive Reinforcement, then, means you’re adding something the horse finds valuable as a reward for getting the response you were looking for. Negative reinforcement includes traditional pressure and release techniques, with an aversive stimulus being removed as a result of the correct behaviour being demonstrated.
Positive punishment involves the addition of an aversive, following an undesired behaviour, and negative
punishment would be the withholding of something good, until the unwanted behaviour stops.
All training approaches and techniques can be explained using learning theory.
When we learn a new technique, it’s important to understand where it fits on this chart, so we are able to use the technique effectively, and possibly even decide if it’s something we’re comfortable applying to our own horses.
Appetitive vs Aversive
In Negative Reinforcement and Positive Punishment, an aversive is used. An aversive is simply something that the horse would rather stop – it doesn’t have to be harsh or painful to be aversive, but it does have to be unpleasant enough to motivate the horse to avoid it. It could be as simple as pressure on a halter or a shift in body position.
In Negative Reinforcement, the aversive is applied to get the desired behaviour, and it’s the removal of the aversive that tells the horse that they have the right answer. For example, you apply pressure to the leadrope towards the horse’s chest, the horse steps back, you remove the pressure to tell the horse that backing up was the answer you were looking for.
In Positive Punishment, the aversive is applied after an undesired behaviour, to make it less likely to be repeated. For example, your horse runs into you, so you correct him with a sharp tug on the lead rope.
Because they’re both involving an aversive, and it’s just the timing of the aversive that is different (before a desired behaviour vs after an undesired behaviour) it’s easy to have some crossover between these two quadrants, whether intentionally or otherwise. Being very aware of your intentions and timing, anytime you choose to use an aversive, will make your communication much more effective. It’s far too easy to find yourself using positive punishment on a horse who has no idea what you’re asking, simply because you didn’t have good timing on the release in your intended negative reinforcement approach.
The same crossover is true for Positive Reinforcement and Negative Punishment. Both use an appetitive, and it’s the timing of the reinforcement that determines a) the effectiveness of the training, and b) whether it’s reinforcing or punishing for the horse.
In Positive Reinforcement, the appetitive (something the horse likes) is given after a desired behaviour, making it more likely to be repeated. For example, the horse stands without touching the human, so the human delivers a treat.
In Negative Punishment, something the horse wants is withheld to punish an unwanted behaviour, and make it less likely to be repeated. For example, the human ignores the horse while he pulls or noses at the treat pouch, and the horse stops because it is no longer rewarding, and THEN uses Positive Reinforcement to reward that choice to move their nose away.
It is important to be careful when using food rewards to use negative punishment sparingly, as the frustration that results can be responsible for many of the unwanted behaviours that those unfamiliar with reward based training can blame on treats.
What your HORSE finds reinforcing
A reinforcement makes a behaviour more likely to be repeated, while a punishment makes it less likely.
But what a horse finds reinforcing and punishing is very individual to the horse.
Food is a primary reinforcer, something the animal needs to survive, so most horses find food rewards reinforcing. But some horses are uncomfortable with taking food from a human hand, or find the training too stressful for them to eat in early stages, and other reinforcers have to be found until they are more comfortable.
The type of food varies as well, and the degree of motivation produced. In some horses, we want quite boring, low value food rewards, or they get too excited. In others, we might need the high octane super valuable food to get the motivation we desire.
In fearful horses, anything that makes them feel safer is reinforcing – if they’re fearful of you, then moving away is probably the most powerful reinforcer in your arsenal at first.
Some horses love scratches, others don’t find touch reinforcing at all – my Ryah is wildly offended when I forget she doesn’t like to be touched and try to pet her after she did a good job. 😉
As training progresses, even a favourite behaviour can be used as reinforcement for a more challenging one.
When I’m driving my horses, I use a free walk as reinforcement regularly – a free walk, where they are stretching through their topline and reaching well under themselves, is a great stretch and we all know that a great stretch just plain feels good. So when I’m working on a skill and they give me a great effort, we immediately do some free walk to let them know that I appreciate the effort.
Our bridge signal (a clicker, YES, or GOOD) also becomes a reinforcer, once it has a strong association with the primary reinforcer of food.
In negative reinforcement, the reinforcer is often the release of pressure, removal of the aversive, or rest.
WTF
You can use an understanding of learning theory to not only train new skills and behaviours, but also to help you reduce or stop behaviours that you no long want your horse to offer.
When you have a behaviour that is unwanted or confusing, the first thing to do is ask WTF? “What’s the Function?”
What’s the function of the behaviour? Why is it reinforcing for the horse?
When we understand why they find it reinforcing, that’s the first step towards avoiding it. Instead we can find a different behaviour to make more reinforcing, change the antecedents or management so the horse no longer needs the behaviour, or as a last resort, use punishment to eliminate the behaviour.
Behaviour: A young colt who bites and climbs on his human.
WTF? Play, roughhousing, energy
Plan: Give him an equine playmate, change his management so he gets more time outside before asking him to behave to human standards, avoid the use of punishment that may be mistaken for engaging in roughhouse play.
Behaviour: Banging the gate at feeding time.
WTF? The horse believes that banging the gate is what makes you bring them food, because of previous experience: “superstitious behaviour”
Plan: Wait until the banging stops to walk towards the horse or deliver the feed, reinforcing NOT banging the gate with food delivery.
Timing
Regardless of which quadrant of learning theory we’re working within, the timing of the reinforcement or punishment is going to determine how quickly our horse is able to understand what we’re asking of them and learn to repeat the behaviour.
If we’re using positive reinforcement, then the timing of the bridge signal and reward is the critical component. It’s easy to accidentally reinforce the wrong thing, which usually isn’t a huge deal (I’ve had many happy accidents, like Bentley offering a “wave” while on his pedestal simply because he had one front foot in the air when I reinforced him, but it is also often the reason that people give up on using positive reinforcement, if their early efforts and poor timing accidentally teaches the horse to pull on their pockets or get overly excited.
The good news is that when we’re using positive reinforcement, the horse has a good feeling about the training as long as we are generous with our food rewards. So if we make a mistake, it’s usually simply a matter of changing the antecedents and figuring out a way to reinforce the “right” behaviour until the horse understands.
We do need to be careful to be aware that every time we don’t give a food reward when the horse expects it, that we are using Negative Punishment. That’s not a bad thing, necessarily, but it can cause a problem if we’re not aware of what we’re doing. If we hold out for the perfect response before we feed, we can easily cause extinction in the behaviour, frustration in our learner, and then probably in ourselves. It we aren’t getting the response we expected, it’s much better to reward the effort and then regroup and ask for a smaller, more accurate step that we can reward with enthusiasm, and build back up gradually to make sure the horse understands what we’re asking. Horses are, as always, individuals, and while some are resilient and love to solve a problem and figure out how to earn the treat, and would do fine with negative punishment as a part of the training process, others are not that confident and are easily discouraged. We need to be very aware of when we are withholding, and how it’s affecting our horse.
While the timing of positive reinforcement is an art, for sure, I think it’s more common to see people have poor timing when using Negative Reinforcement, to the detriment of their horse.
If you’re using negative reinforcement – even minimally aversive, non-escalating, like lead rope pressure or a touch of your hand – and you don’t remove it as soon as your horse shows any effort, then your communication isn’t effective.
You use lead rope pressure to ask your horse to back up. They rock their weight back, you release, and the next time you’ll probably get a step back, you release, and the next time you’ll get more than one step, because you have clearly explained to the horse that backing up is the way to remove the pressure on their halter.
Unfortunately, we humans tend to get greedy, and I think it’s even more common with Miniature Horses, do to their small size. What often happens instead is, we apply halter pressure, the horse rocks back, we think, “yay, it’s working!” and keep applying pressure, and the horse might take a step back but at some point they’re going to start thinking that backing up wasn’t the answer we were looking for after all, and we’re going to either activate their opposition reflex, and they’ll just answer the pressure by pushing back at us, or they’re going to start trying something to see if that makes us stop the pressure, and the “something else” might quickly escalate to things like rearing and crashing into us.
Again, it’s easy to accidentally cross over into the Positive Punishment quadrant from Negative Reinforcement (because they both use aversives), and by continuing to apply pressure without reinforcing the small efforts, you’ll wind up with a confused horse who has done the behaviour you wanted, but is now being punished (aversive applied after the behaviour) and no long wants to repeat it.
On of the reasons that I now choose to keep my training centered in the Positive Reinforcement quadrant whenever possible is that if I make a mistake then my horse doesn’t learn what I was hoping, but they likely don’t have any bad feelings about me and the training. They still got to play and learn and earn food, so they’re probably pretty happy. When things go wrong in other quadrants, it’s a much different emotional experience for the horse.