We talked about bribery before and what that produces in a horse versus using positive reinforcement (+R). This time we want to talk about the learning theory behind bribery versus +R.
Bribery – Learning Theory
Dictionary definition: The bribe is the gift bestowed to influence the recipient’s conduct. It may be any money, good, right in action, property, preferment, privilege, emolument, object of value, advantage, or merely a promise.
In other words, you use the reinforcer to elicit the behaviour. How is this different to +R ?
+R is where we reinforce a behaviour that is already happening.
Bribery is about using the reinforcer to get the behaviour to happen.
The difference is huge. In one we capture something that the horse is already doing, in the other we make it happen with the promise of a reinforcer.
When we capture behaviour then we are really only using +R. We wait for behaviours and then click and reinforce. Once we have that happening reliably we can start to shape the behaviour. For example, picking up feet. We can wait for the horse to move a foot, click and reinforce. When they move it again, click and reinforce and so on until they understand it is about moving that foot.
Picking up feet can also be trained using bribery. We can use the food infront of the horses nose to get the horse to move so that we get the foot to move. Once the foot moves we can click and reinforce. However, at this stage we have to ask what the horse might think they are being clicked for….following food or moving a foot ? With +R, the food is not being used to elicit anything and so the horse has a better chance of figuring out what they are being clicked for…we are setting them up to problem solve.
It used to be said that the click reinforces the behaviour that is happening at that moment in time. However, as the understanding of learning theory and clicker training has evolved and the behavioural analysts have studied the effects of clicker training more, we now know that the click actually reinforces ALL the behavior that came before it. We need to bear this in mind as we train, whether that is eliciting the movement required or waiting for it to happen. We need to understand what it is we are reinforcing.
Understanding the Behaviour
The difference with using bribery or +R to train behaviour is the horses understanding of the behaviour.
When we capture/shape the behaviour, the horse is learning to problem solve to work out what it is we are asking for and what we are reinforcing. When we bribe a behaviour, all the horse has to do is to follow the food. As such, the result of bribery is that the handler owns the behaviour and the horse cannot replicate that on their own as they have not worked out how to do the behaviour themselves. When we use +R what we find is that the horse can repeat the behaviour without guidance. What we also see is that the horses then practice this behaviour on their own.
Dictionary definition: Something that tempts or attracts with the promise of pleasure or reward
Interestingly, luring is very similar to bribery.
Luring is something that is used in the dog training world. The food is used to guide the dog to perform behaviour. It’s a quick way to get the dog moving and capture behaviour. One of the rules with luring is that you have to stop using the treats in your hand very quickly. The rough number is 3 times (this will vary from one dog to another depending on how quickly they pick up the behaviour).
Once you have the behaviour happening 2 or 3 times, you have now taught the dog to follow your hand. So when you give that hand cue (minus the food) the dog will follow the cue. As such, with luring, we are teaching the cue for the behaviour at the same time as we teach the behaviour.
Horses versus Dogs
With horses, luring is not something I use nor encourage. To start with, there is a big difference between luring a dog and a +550kg horse. One of my horses is food orientated and I can tell you how scary that prospect is to think about luring him.
If you have a horse who can get anxious or grabby about food and you are asking them to follow food, you can create some very unwanted reactions which you will have to deal with as a result. Even if you have a horse who is not particularly food orientated, you can get use luring and get energy levels up so high that they can quickly be out of control before you know it.
With luring, the food becomes a moving target…not something horses have to deal with to eat. Their food is stationary. So having to chase their food is probably not fun, it is certainly not instinctive. However, dogs are born to chase and kill their food. So following their food is not stressful, it’s fun, that is why we can use it to encourage them to learn new behaviours. We can also use it to play with toys (the kill instinct) to reinforce other behaviours and more.
We can also use play with horses, we just need to be cautious about using play that involves chase, and using play around humans. Dogs jumping on us out of stress from overfacing them with chasing food or getting them over excited is a whole different ball game than a horse jumping on us for the very same reason.
One of the common questions on the forums is about how to manage horses who get over-excited when clicker training. With luring we could easily have a horse like that AND be asking them to chase food, or as a result of chasing food.
When we get horses to chase things we can be bringing out a behaviour that we are not able to manage and that raises unpleasant emotions in the horse. When horses chase something they are chasing it to defend something; to attack the other thing and inflict injury. In geldings this is the horses’ instinct to protect the herd and with the mares this would be the instinct to protect a foal. Both very strong emotions and not ones that the horse will perform without good reason. The emotions that come with that behaviour are not the emotions we want the horse to be feeling when we train. They are not conducive to good training.
When we train, we have many options open to us for how we achieve the behaviours we want, and it is good to know how to train for behaviours in different ways. The more ways we can train for behaviours the more rounded the training will be. However, we also need to be skilled enough to know when to use each method. The golden rule being; if in doubt stick with +R for behaviours that are already happening.
Something else to keep in mind as we train is something that Alexandra Kurland says frequently; “Even though we are training using +R does not automatically mean the animal is having a positive experience”. A topic for another blog.