Positive reinforcement – taking the reins for equine training

I remember it so well. I was reading my first book on dog behaviour and was impressed – the trainer was pictured happily walking a collection of dogs off-lead.

The book was endorsed by an eminent natural horsemanship trainer I rated highly and contained practical advice about how you could ensure a dog behaved as desired by humans, by following certain rules of wild canid behaviour.

You simply had to place yourself in the dominant position within your “pack” and behave in a manner in keeping with being the leader – by eating first, leaving the house first, greeting a dog only on your terms and never permitting the dog to overtake you on a walk. After all, dogs were descended from wolves and wolves behave like this.

This was a while ago. It was before I had owned a dog or became an equine behaviourist.

Doubts started to creep in as I became friendly with knowledgeable, dog-owning horse owners who pointed out flaws in these so-called dominance theories.

I noticed a lot of people on internet forums liked the so-called “natural horsemanship”, but simultaneously disliked the dog version.

People would use reward-based training for their dogs, but believe, erroneously, it would encourage their horses to bite. People would use pressure halters or gum lines on horses, but reject the use of choke chains and other aversive collars for dogs.

I failed to understand why this cognitive dissonance existed, but I put it to one side and didn’t think more about it.

As my interest in equine behaviour grew, I was on the lookout for quality books about the subject. And it just so happened a number of the best books on animal behaviour and learning were written about dogs.

It turned out dogs were not so much descended from wolves, as dogs and wolves having descended from a common ancestor (Bradshaw, 2011). This isn’t semantic; it means no wild animals exist today that can tell us how our domestic dogs should behave “naturally”.

I also learned, even if we assume this common ancestor was sufficiently wolf-like, today’s wolves do not actually behave in the way proponents of dominance models have claimed – the so-called dominant pair are simply the parents of the offspring comprising the rest of the pack.

Dogs, on the other hand, seem to have evolved to fill a niche of mutualism, co-existing alongside humans. Less pack behaviour and more opportunistic individual (Coppinger and Coppinger, 2002).

Different perspective

From the perspective of an equine professional, but lay dog owner, I look into the dog world with wonder.

Dog people really seem to “get it”. The Dog Welfare Campaign fought to keep an eminent and infamous dog trainer out of the UK, on account of his abusive training techniques.

This is an amazing success story for dogs. Although the trainer still visited the UK, he was challenged in a number of mainstream television interviews, causing a significant reduction in his ticket sales.

When trainers tell us dogs are pack animals and look to us as the pack leader, enough dog people are out there telling us this is rubbish on numerous counts.

Aside from the ethology of wolves and dogs, there is no evidence to suggest dogs look to us as any sort of dog, leader or otherwise. Treating them aversively may obtain compliance, but it also causes them to fear us. The emperor’s lack of clothes has been well and truly exposed.

But this is where I get confused. A lot of dog people are also horse people and many of these believe horses behave badly out of dominance.

Apparently, all we have to do if we want to overcome the problem of being dominated in our little herd of two is become the leader, and this means using a series of aversive techniques – supposedly because that is what the herd leader would do – to impose our will on the horse, without considering what the horse may think about it all.

After all, we can convince ourselves horses have been relieved of the burden of keeping our herd safe.

Not only have horse people failed to coordinate a campaign to prevent aversive horse trainers visiting the UK, we have allowed ourselves to become inundated with them.

We welcome them with open arms; be it a novice horse owner or the royal family, we continue to invite trainers here and apply their supposedly “natural”, but undisputedly punitive, techniques to our unsuspecting horses (Figure 1).

We repeatedly watch slight variations on the theme; sending horses into flight around pens or causing them to come into contact with narrow rope halters at speed to “put the pressure on themselves” are among favoured techniques trainers use to impart their message of “being the alpha mare”.

Just as in the dog world, it turns out we have been sold a mythical version of equine ethology and there is no evidence to suggest horses see us as horses.

Horses are herd animals and generally enjoy peace (Figure 2). In the wild, they typically form herds comprising family groups where the dam is the so-called “leader” and her offspring make up the rest of the herd.

The leader hasn’t so much as demanded the role, it just has knowledge of good locations for key resources. The stallion may assume some part of the leadership, but as more of a protective role, driving the herd to safety (McGreevy, 2004).

Some herds have additional stallions, apparently adding to the stability of the herd. Of course, there are also stallion-only, or bachelor, groups and, predictably, these are less stable with more agonistic interactions, but still with no sign of the alpha mare so beloved by certain trainers.

Science has attempted to identify dominance hierarchies, with little success, unless the resources are restricted some way (such as Rubenstein, 1981). When disputes occur between two individuals it is not so much because one of them is dominant over the other, but because one of them values a particular resource (such as water or food) enough to challenge the other horse.

Horses can engage in a repertoire of threatening behaviours to avoid the actual fight. The “winner” may obtain the resource, but does not outrank the “loser”. More generally there is a give and take.

It is far from the simplistic leader-follower model so many trainers would have us believe (Berger, 1986).

When you have a horse in a domestic setting, we automatically restrict all its resources. We interfere with its choice of companions, even if in a stable and permanent herd, never mind if the companions are ridden, stabled at night or switched around. We restrict space and grazing by having too many horses on insufficient land. We place water troughs at the edges or corners of fields where it is difficult for social drinking to take place.

The dominance hierarchies observed in so many domestic herds are more pathological than representative of true equine behaviour. Then we wonder why horses develop behaviour problems.

When we want to deal with these behaviour problems, pretending to be the “herd leader” must utterly confuse the horse, compounding its difficulties.

Again, there is no evidence to suggest the horse recognises our behaviour emulates that of a horse. Forcing it to engage in a series of dominance-reduction exercises may obtain compliance, but is as unlikely to result in the horse’s problems being solved as in the case of dogs.

Training seen by the horse as aversive will not result in it viewing you as leader, dominant or trustworthy; it will just learn to fear you and to stay out of trouble and fighting back if really pushed over its limits.

Embracing a more positive approach

While die-hard proponents of dominance-reduction dog training techniques remain, it seems they are becoming increasing marginalised.

The dog world has embraced positive reinforcement. It is relatively straightforward to implement and is fairly easy for a novice dog owner to recognise. For example, a “recall” with a treat is more likely to be successful than a smack for running off in the first place. Horses, on the other hand, are ridden.

We have centuries of riding techniques at our disposal, all based on punishment and negative reinforcement, and the odd big loud pat, which we often claim is positive reinforcement, but really doesn’t seem pleasurable to the horse.

Horses are big and we don’t like to admit they often scare us. So they must do as they are told and we claim, euphemistically, this is necessary for the sake of our safety, a somewhat foolish approach if we are going to insist a frightened animal suppresses its emotions.

If horses don’t do as they’re told, they show us up in front of friends, livery yard owners and show audiences. We can’t have that.

We also tend to skip a lot of the early training they need and expect them to just get on with it. There are plenty of punitive corrections we can apply later.

Mainstream horse people do not really embrace the idea of behaviourists and applying science to what are typically misunderstood to be training issues.

We like to think we give horses “the best care”, anthropomorphically meaning a deep bed, cosy rug and three meals a day, and think they should be grateful to us. And we think we “need” to have our horse at the yard with the floodlit school, rather than in the herd in the field down the road with no human-centric facilities.

We don’t like to think their ethological needs are not being met. We don’t like to think we need help from someone in “a white lab coat” who’s going to talk operant conditioning at us.

We want to “love” our horses instead, while remaining blind to the feelings they may have for us.

But things are starting to change. Positive reinforcement has been creeping into the equine scene in the UK over the past 20 years, via the emergence of equine behaviourists and/or clicker trainers.

In many ways it heralds the dawn of a new equine world. We are starting to listen to the reasons why horses may behave in ways we don’t like.

We assess whether pain, lack of turnout, lack of company and/or an inappropriate grain/sugar-based diet are underlying causes of behavioural problems. Sometimes we even reward actions we like instead of punishing those we dislike.

But we need to be careful. There is already a tendency to assume the horse is automatically having a good time because the training method is based on rewards.

If ethological needs have not been met, or if the horse is being asked to perform when he feels frightened, there will be a distinct absence of positive experience for him. The same is true of dogs.

In a funny sort of way we are converging, but for the wrong reasons. We need to step away from training methods and think more about the foundations of equine and canine behaviour. Not the ridiculously over-interpreted, anthropomorphic version of equine and canine ethology, but the real thing.

We already know the real versions, though, admittedly, they are possibly less well marketed outside academia. This may mean acknowledging some of the demands we make of our animals are unreasonable and should be reconsidered.

This may mean examining our motivation for accepting aversive methods for horses that we would reject for dogs.

It may mean recognising the methods we thought were benign are upsetting our animals more than we care to admit. If this means there is less need for overpriced abusive trainers interfering with our animals, then it’s for the better.