A common criticism of those who train horses using positive reinforcement is that we are so busy discussing behavioural theory that we do not do anything practical with our horses, just a few “tricks”. Or that our training is so constrained by theory that there is no “feel”. Or that what little practical work we do with our horses takes so long it is not viable for most people. Or that we have dangerous horses who gallop into busy roads and leave us waiting desperately for them to stop so we can click and treat.
I’m not being facetious, I have been accused of all these things and I would argue that none of them is true. So what do we do with our horses?
Most of us learn to use positive reinforcement via clicker training. And when starting clicker training it is true that most of us start with simple targeting exercises that may be perceived as “just tricks” by the uninitiated. But targeting is considerably more than just a trick. It involves the horse spontaneously touching a novel object in order to earn a treat. The handler clicks at the exact moment the horse performs the correct behaviour and this helps the horse to understand which behaviour has earned the reward. In order to succeed, via a certain amount of trial and error, the horse must overcome any fear or wariness of the target, it must inhibit any other behaviours such as mugging or biting and it must make a choice to act autonomously. The horse also starts to associate us and our training with good things happening. So even in the early stages of clicker training, we are using the clicker to help the horse develop in confidence, self-control and personal growth, as well as potentially helping to improve our relationship. Not bad for a few minutes’ work.
A free-shaping session such as this (i.e. using pure positive reinforcement without cues or lures) can be particularly valuable for a horse who is reluctant to offer behaviours as a result of previous aversive training. It provides a safe environment where mistakes are tolerated and not corrected. The horse can learn to make choices, secure in the knowledge that there will be no negative consequence of choosing the wrong answer. Free-shaping can therefore be an extremely valuable tool in the rehabilitation of mistreated horses, with very strong analogies with human counselling. An acute level of “feel” is crucial, taking this approach well beyond the crude “stimulus-response” training of the 1950’s behaviourism movement.
But for the average horse-owner who is not trying to rehab a rescue case…..
Clicker training can be a great tool for solving minor problems. On one livery yard I had to take my horse across a dairy pasture in order to reach his field. All the horses would dive for the grass and we would struggle across, trying unsuccessfully to hold their heads up. I thought it would be a nice clicker exercise and used shaping to teach my horse that it was OK to graze when he heard the click. Initially I would click every couple of strides *well before* he tried to dive for the grass. He started to wait for the click because he knew he was then allowed to have grass. Gradually we increased the number of strides before the click. It wasn’t long before we could cross the dairy pasture before grazing â€“ unlike all the other horses who continued to dive for the grass. I like this example as it illustrates nicely that, although clicker training and shaping may initially appear to be long-winded, they actually save time and solve problems more quickly in the long-term because we are appealing to the horse’s choices rather than fighting them.
Some clicker trainers choose to have a clicker with them at all times so as to “capture” any behaviour they like at any time. Thus clicker training can be used alongside any general handling or riding that people do. For various reasons (and a whole new article in itself), I prefer to reserve clicker training for well-defined clicker sessions but those sessions might specifically be for teaching behaviours such as picking up feet, loading, leading, standing still or learning to move away from light physical pressure. Most commonly I use clicker training for free-shaping over, under, through or around obstacles, picking up a toy or pushing a football for increasing confidence, patience and enhancing a relationship based on mutual trust and choice. I also use it as a way to give my horse scratches on his itchy spots without him demanding too “emphatically” – he will spontaneously back away from me to “ask” for a scratch which is much safer than his previous barging.
Perhaps another key point is not so much what I do as what I do not do. I try to be aware of any inadvertent reinforcement I might be giving my horse which encourages him to behave in ways I see as undesirable. I take note of any behaviours he gives me and, instead of trying to stop them happening, I try to ignore them* and learn the circumstances under which they arise. This takes me to the root cause of the behaviours and so I can remove the cause, rather than worry about the behaviour which typically then disappears of its own accord. Ignoring unwanted behaviours is an essential part of training with positive reinforcement and is perhaps one we tend to over-look when we are thinking about “what to train”. Learning to just sit and observe is difficult, particularly if we perceive that our safety is at risk, but the more I trust in the horse’s innate cooperative nature, the more I can avoid confrontation, increasing both our safety and our mutual trust yet further.
When not engaging in a clicker session I am happy to use mild pressure to make requests of my horse, particularly when riding. But that does not stop me from using the basic principles of learning theory – I am careful to release pressure with good timing and I try to keep the pressure constant so that the horse has a chance to learn how to release it. And, perhaps most crucially, I continue to use shaping. Shaping – i.e. the breaking down of any task into its tiniest component steps – is arguably the factor that is the difference between keeping safe and becoming a liability. If I do not want to exert excessive pressure on my horse in order to keep us safe then I need to have completed sufficient early training that excessive pressure would never be required. It is shaping that almost guarantees that we will not have a dangerous horse who gallops into traffic because we would have never put him in a situation like that – we would have devised a shaping plan with an end goal of “riding safely in traffic” and broken the task into many training steps. There may be the odd rare occasion for which we cannot prepare, but the more we use shaping and a non-confrontational approach, the less we find that our safety is compromised.
(* it may sometimes be necessary to extract myself as quickly and as safely as possible, perhaps resorting to aversives if need be – but this would be a one-off situation into which I would avoid getting again without additional prior shaping/training)