back to previous page

How horses learn: classical conditioning by Sara Lewis

One of the ways people and horses learn is by association, that is, we learn to associate one event with another. Some of these associations are presented to us as part of the environment (they are just there) and we learn about them by interacting with the environment. So for example a nettle sting causes pain, so we quickly learn that touching nettles is painful. For a horse an electric fence causes pain, so they learn to avoid touching the electric fence. Similarly we learn that sugar is sweet or for a horse that pony nuts taste good. These are very basic associations and you will notice that they are focussed on physiological states (hunger, thirst, fear, pain, anger etc). These associations, once learnt, influence our behaviour, we avoid nettles, we seek out sweet foods, horses also seek out food and avoid painful situations. Learning like this is a basic survival mechanism: eating that food made me feel ill, I will avoid that food; Touching that fence made me feel pain, I will avoid touching that fence; Eating that food made me feel good, I will seek out that food. These associations are formed at a physiological level and may not be easily available to our rational mind. And when they are available can't necessarily be easily overridden, as demonstrated by phobias: I 'know' that spider can't hurt me and still I feel sick with fear when I see it. We form associations like this all the time, in an involuntary way, and our behaviour is informed by our understanding of these associations. For the horse too various involuntarily formed associations between features of the environment and his physiological state inform his 'world view' as it were. 

The second thing to notice is that these associations can become linked to other closely associated features of the environment.  So for the horse for example food often comes in a bucket, so the response to seeing a bucket becomes to seek to stick their nose in it, in anticipation of eating. Now the interesting point is that sometimes the original physiological pairing is so overlaid by other associations that it is hard to see what is going on. Let's suppose we have a horse that is consistently really difficult to saddle. We could say he's a naughty horse, trying it on, or we could try a learning analysis. We could hypothesise that at one time the horse had a sore back, and that the horse quickly learnt to associate being ridden with an increase in pain. It may than have associated the act of being saddled, with being ridden, with an increase in pain, since whenever it was saddled it was ridden and whenever it was ridden it was in great pain. The horse wants to avoid the pain, so it tries to avoid the riding, and in due course by association the saddle (perhaps by rearing up, or lying down, or kicking out) which is now associated with pain. Let us further suppose that the horse's back was cured of pain, and theoretically riding should no longer cause pain, however the horse still reacts badly when it sees the saddle, why is this? This is because the avoiding the saddle behaviour (rearing up, kicking out or whatever) has become a conditioned response to seeing a saddle. It is important to appreciate that the horse has no control over this, it is like a reflex reaction. My body sees a saddle, my body anticipates pain, my body takes avoiding action.  

Looked at in this light we can see the horse's behaviour in the present as stemming from  a series of historical associations that no longer hold true, however the horse doesn't know that. Many people might work from their own perspective and assume since I know this isn't going to hurt you I am going to make you let me put it on (A kind of 'get over it horse!' attitude). In many cases this won't cure the problem as by the time the horse is saddled it is in a state of fear or stress and so the saddle continues to be associated with a state the horse would rather avoid, so inadvertently we are continuing to create unpleasant associations with the saddle. However, if we take a learning perspective we might say, I don't know how the horse came to form this association, what I do know is that I want him to start to form a different set of associations. I want him to start to associate the saddle and being saddled with nice feelings and sensations. How might we do this? There are a number of ways we might start. We might bring the saddle as close to the horse as we can without evoking the avoidance response and then pat and pet the horse, feed him, groom him etc. and then take the saddle away, gradually moving closer at each exposure. We might see if he will accept something really small on his back, say a handkerchief and then gradually built up from there. We might ride without a saddle, so that the association between the saddle and riding is broken, and the horse begins to learn that being ridden doesn't induce soreness. In all these ways we can begin to help the horse to learn that his associations, and so reactions, are outdated: in other words we can help him learn that it is safe to change his behaviour. 

Next time we'll look at associations formed as the horse learns how to make things happen.

back to top