I recently received the following question from Sharon B in Taranaki  

"With the requests for a particular Action from your horse do we give the horse a voice command to accompany the Action we want the horse to take, or is it a random action that we reinforce i.e. 'look away'.  e.g. do we say "Look Away" when we want the horse to do that or do we just reinforce the looking away action?"

It's is a common question from those new to clicker training. It is almost inconceivable to many novices that we can create a behaviour apparently out of thin air but the only alternative appears, at first, to be that we clicker trainers must be hanging around for hours on end waiting for a random action from our horses.  

With negative reinforcement training we are ingrained in the dominance based and\or pressure based systems which rely on us physically directing our horses.    We know that we must see the behaviour before we can Click! and reward it.  However, in the initial period of introduction to positive reinforcement we often find it hard to think how we might get our horse to do anything if we do not use pressure or punishment or at least some command.

So let’s address this.  Even tied quietly, the average horse is looking around him, swishing his tail, changing the angle of his ears, sniffing things, sometimes nibbling things and often changing weight distribution on his legs, or perhaps even taking a step or two to relieve the boredom or to check out what's going on in different directions around him.  Any of these actions can be reinforced in order to encourage them and turn them into what you want.

This is exactly what we do with the foundation exercise 'look away'.  We know how a horse normally behaves and we simply reinforce the bit that we want.  In this case, we know that a horse will look around him and so we begin to reinforce this when it happens and end up with a behaviour that counter-conditions* the horse in order to produce good food manners. 

Here, then, we have created a behaviour without commanding the horse to perform what we wanted or physically manipulating him in any way.  This can be a revelation to new trainers.  It is not however, random or accidental.  We can do this because we understand how horses behave and are willing to pay attention, to be observant of our horses and see individual actions. By being observant we can also be predictive.  We can know what our horse is likely to do and use our clicker to precisely capture the piece of behaviour we want, as soon as it does occur.   

This seems simple but it is a significant change in the way you may currently think and can take some time and practice to get used to.  At first the seeming randomness puts many novice trainers on the back foot.  It seems too long a process, too hard too….. something that they can’t quite put a finger on to work, to be useful.  Yet, this ability to see what the horse already does, to wait, to watch, to understand how the horse naturally behaves is perhaps one of the greatest gifts that clicker training can give the trainer. 

This is why we try to start novices clicker training using simple exercises that use shaping and free shaping (starting a behaviour from ‘nothing’) instead of simply adding the clicker on top of negative reinforcement.  It enables the change in mindset required to really take advantage of positive reinforcement instead of muddling it with traditional command based training.

We overcome our human trait of commanding and in effect, doing everything for the horse, in favour of directing the outcome – a far more subtle approach- and allowing the horse to do the work himself.  Later, if we perhaps do use the clicker in conjunction with negative reinforcement we are much more thoughtful about it and can use greater finesse than if we had not been through this journey of discovery and changed our mindset.

© Horse Play 2006

* counter-conditioning is a term that means we teach a behaviour that is the opposite to (or at least cannot be performed at the same time as) a behaviour already in existence.  In the case of 'look away' the normal behaviour for a horse would be to look toward or even move toward a known food source - he gets rewarded for doing the opposite, looking away.  He cannot both look away and look at or move toward the food - counter conditioning.