Are you using positive reinforcement or negative reinforcement with a cookie on top?
Picture by Janneke Koekhoven
Some trainers may, in an attempt at using positive reinforcement, add a click and a treat to their existing training approach. And while this is a valiant effort, it is important to remember that the presence of a clicker and some food does not necessarily mean that the horse is experiencing positive reinforcement. A good way to determine if positive reinforcement is being used is to try and answer the following questions:
-
What happens to the behaviour if clicking and treating is removed from the session? If the horse is working for the food rather than the relief from aversive pressure, he should lose motivation and the behaviour should degrade in the absence of reinforcement. There is of course exception to this rule as the horse may be motivated by something outside the handler’s control. For example, many horses are keen to get back to get back home without needing their rider to nag them with their legs (negative reinforcement) or to tangibly reinforce their walk with food rewards (positive reinforcement) because they want to see they herd again.
-
What does the handler do if the horse does not respond to the cue? If your answer involves escalating the “cue” by repetition (e.g. repeatedly squeezing the horse’s flanks until the horse respond) or strength (e.g escalating to a kick or smack of the whip), negative reinforcement is being used, not positive reinforcement.
-
How is the behaviour trained? Looking at the final behavioural product
alone it can be hard to determine if something has been trained using
positive reinforcement or negative reinforcement + cookie because some
clicker trainers use non-aversive tactile cues that look traditional. Therefore, it is best to look at the training process. Specifically, how did the handler initially get the wanted behaviour. If it involved applying any sort of aversive pressure (may it be tactile or just the threat of an aversive touch with a rope or whip) then regardless of the presence of food, it is not an example of positive reinforcement training.
Aversive pressure vs tactile cue
Due to popular trainers using the euphemism “pressure and release” to describe negative reinforcement-based training, pressure has become a dirty word, making it difficult to imagine non-aversive tactile cues possible. It is worth stating that while lots of “pressures” in the horse world are aversive some may be appetitive, for example the pressure created when giving your horse scratches and some may be neutral, for example the pressure created by a rug.
Whether a “pressure” is aversive or not is determined by whether it trigger:
-
An unconditioned response of fear, discomfort or pain. This mean that without prior learning the horse finds the “pressure” unpleasant and will naturally seek to avoid it. For example, a whip blow.
-
A conditioned response of fear. This mean the horse was taught to perceive a “pressure”, that he may have found to be neutral or even appetitive in the past, as aversive through classical conditioning. For example, a small leg squeeze that triggered no specific response prior training became aversive by being associated with a kick or whip blow.
When you see a positive reinforcement-based trainer use a tactile cue, this trainer would have first done two things. 1. Make sure that the cue he wants to use does not trigger any specific behaviour and 2. Teach the behaviour prior to introducing the cue. If this is not the case odds are this trainer uses negative reinforcement with a cookie on top.
Positive reinforcement-based trainers that work with horses rarely use tactile cues as they understand most horses, they come across have a past of traditional, aversive-based handling and riding. As a result, these horses find most touches aversive, especially in the context of riding. There is the option to counter-condition the horse to no longer find certain pressures aversive, but doing so may not always be useful or necessary, and will take significant additional training time when a word or visual cue may be just as good.
