|Posted by Hillydale Equine Training and Sales on July 11, 2013 at 8:55 PM|
If I had a dollar for every time I heard or read this statement I’d have post and rail fencing and possibly an indoor arena! Many trainers whom I personally admire and respect confidently assert that using food to modify horse behaviour is simply bribery. It’s a commonly expressed view of many guru natural horsemanship trainers. Back Brannaman, Pat Parelli and Clinton Anderson have all stated that horses trained with food don’t "respect" their trainers. In fact Pat Parelli has written in his seminal work Natural Horse Man Ship that food rewards don’t work for horses at all.
So are they right- is food a bribe and are training outcomes rewarded with food less durable, reliable or genuine?
In a word-no. These views aren’t supported by the findings of a number of studies of equine learning and cognition and they demonstrate a significant misunderstanding of how horses(and the majority of animals) actually learn.
Like all animals, horses are highly motivated to get accessto food. In fact many of their key behaviours are geared to finding, collecting and digesting food. Depending on the environment in which they live, horses will employ a variety of strategies to access food, from pawing through snow to reach the grass underneath, harvesting berries from trees or opening gates to get into feed sheds. These behaviours have to be learned and if they result in food they will be remembered and repeated. This is known as “reinforcement”-whereby a behaviour is strengthened or repeated because it results in a good outcome-for example food! There are other forms of reinforcement which rely on the removal of something unpleasant such as pressure. More about that below.
Calling food rewards bribery reflects more about our beliefs about what horses “owe” us than horse cognition. It demonstrates that we believe horses should do what we request of them- such as enter a float or stand still to be mounted simply because it’s a human who is doing the asking and that human is owed obedience. It assumes that horses “know”what they should be doing and that they can be corruptly influenced to do it. In human society, bribes are paid to people who should be providing a service but won’t unless they’re paid that something extra. Bribes are paid to policemen to look the other way, to public servants to speed things up and to politicians to pull strings to gain advantages not available to ordinary people.
By calling food rewards a bribe, we’re making another assumption, that when horses are compliant it’s because they “want” to respond as we wish, they are “willing” or it’s our skill as a trainer that has gotten such a large animal submissive to our will. Yet in fact, it’s just that we’ve used a different kind of reward- the removal of pressure. Training techniques that don’t use food usually rely on the application of an unpleasant pressure to the horse’s body, the removal of which is rewarding. This is known as negative reinforcement because the removal of something (pressure) is rewarding (reinforcing). Pressure-release is the most common form of negative reinforcement and it underpins almost all human-horse interactions and by far the majority of training systems, techniques and approaches. Whatever the label attached to the training method or philosophy, the use of pressure and its removal will be its basis. But the reason the release of the pressure is rewarding is because the pressure itself is unpleasant. So to give our horses the reward of release we first have to do something that makes them feel uncomfortable and may even cause them pain. Yet we don’t call the release of pressure a bribe.
Horses are not moral beings- they don’t have insight in to their or our behaviour and they don’t “choose” not to comply with our reasonable requests- they simply behave in ways that maximise good outcomes and minimise or avoid bad outcomes. Two common good outcomes for horses are escaping/avoiding pressure and receiving food. They are two sides of the same coin and they both involve rewards. It’s just that one requires something unpleasant to happen first to create the reward and the other doesn’t.
What does the science say?
Studies comparing food rewards (positive reinforcement) with pressure-release (negative reinforcement) have demonstrated that horses trained with food reinforcement are more likely to voluntarily interact with their handlers, show more interest in people in general and appear more willing to engage in training interactions than horses trained with pressure- release [4-6]. Australian trainer Georgia Bruce has used food reward training to teach her horses to perform piaffe at liberty and undersaddle, paint pictures using a paint brush, lie down, rear, play soccer and performflying changes to list just a few. Food reward training has also been found to be more effective in gentling abused horses than a pressure-release technique.
Horses can even be trained to load onto a float or trailer using food rewards, without any pressure-release at all. One study compared using target training (where the horses were food rewarded for following a target) to get difficult loaders onto a float, with a standard pressure-release method. Although there was no difference in the time taken or the number of horses who finally loaded, the food rewarded horses had significantly lower heart rates and performed significantly less fearful behaviours than the pressure-release trained horses. For a good step by step description of using target training for float loading see the current issue (July 2013) of Horses and People magazine.
Food reward training is the primary method used to train marine mammals such as killer whales, seals and dolphins and it is increasingly used in captive animal handling to train animals as diverse as tigers, chimps,rhinos and otters to submit to routine husbandry and vet procedures. It hasalso been used to train young elephants to be ridden without the use of chains or restraints. Trainers of these animals never refer to food as bribery.
There are some horse trainers who believe that horses should only be trained using food rewards, however unless we never use any kind ofequipment on our horses it’s pretty much impossible to avoid using some form of negative reinforcement (pressure- release). Used inccorectly, food reward training can result in some unwanted behaviour such as mugging which can become a problem if not addressed. Noted equine behaviour scientists, Dr Andrew McClean and Professor Paul McGreevy have observed that by far the majority of unwanted and dangerous horse behaviour such as bucking, rearing, baulking, kicking and bolting result from incorrectly applied negative reinforcement. They also note that because negative reinforcement relies on the deliberate application of an unpleasant sensation just so we can remove it to create the reward, that the potential welfare impacts of getting it wrong can be catastrophic for the horse . On the whole, mistakes made in food reward training simply mean you dole out more food than planned which is unlike to compromise the horse's welfare and it would be extremely rare to end up being bucked off or the horse trial rearing or bolting if you get your cues mixed up when using food rewards.
There are training situations where food rewards are of limited value because the horse isn’t motivated to eat. We often see this when horses are fearful or anxious. In those situations the horse’s focus is on survival, which in evolutionary terms has been achieved by fleeing whatever is scary in the environment. Most horses, given a choice between eating and running away from a lion will choose the running. Substitute a float for a lion and result can be the same. In those instances,pressure-release techniques can be the most efficient means of achieving a training outcome. However, often times the reason the horse is frightened is because we haven’t taken the time to familiarise it with the stimulus or situation or because it is confused about the task. In these situations providing food rewards can assist to calm the horse down by creating a pleasant association between the scary situation and a good outcome- food. I frequently use a combination of pressure-release and food rewards when training nervous horses to load and then stay on floats. Most horses quickly come to associate the float with getting food and they become so focussed on learning how to get the food that they in effect forget about the float- a process called habituation. I’ve often seen the behavioural signs of fear- snorting, high head carriage, rolled eyes, etc quickly disappear once the pattern between my cue, behaviour and food reward becomes clear to the horse. The same result can also occur when pressure-release is used correctly and I’ve often used pressure-release without food rewards to train horses to load. In my personal experience I've found that the food reward trained horses calm down more quickly and "volunteer" to enter the float more readily.
Novice food reward trainers often comment on the huge improvement in their horse’s motivation and willingness to try new behaviours compared to before they started using food rewards. This is unsurprising- a tasty treat is all upside and doesn’t rely on creating an unpleasant sensationt that has to be terminated to create the reward.
There is and always will be a place for pressure-release in horse handling and training and I am a daily user of pressure-release techniques as well as food rewards. But its time horse trainers opened their minds to what science and trainers of other animals have known for nearly forty years-food rewards are highly motivating and result in excellent training outcomes. As a reward, food doesn't involve making the animal uncomfortable first and its potential to cause harm or set back training is minimal compared to pressure based techniques.
If food rewards aren’t a bribe then what are they? I like to think of them as wages.
1. Brannaman, B., The Faraway Horses-The Adventures and Wisdom of One of America's Most Renowned Horsemen, 2001, Guilford, CT: Lyons Press.
2. Parelli, P., Natural Horse.Man.Ship, 1993, Fort Worth, TX: Western Horseman.
3. Anderson, C. and M. Kaitcer, Lessons Well Learned: Why my method words for any horse, 2010,Wollombi, NSW: Exisle Publishing.
4. Innes, L. and S. McBride, Negative versus positive reinforcement: an evaluation of training strategies for rehabilitated horses. Applied Animal Behaviour Science, 2008. 112(3/4): p. 357-368.
5. Ferguson, D.L. and J. Rosales-Ruiz, Loading the problem loader: The effects of target training and shaping on trailer-loading behavior of horses. Journal of Applied Behavior Analysis, 2001. 34(4): p. 409-423.
6. Sankey, C., et al., Positive interactions lead to lasting positive memories in horses, Equus caballus.Animal Behaviour, 2010. 79(4): p.869-875.
7. Bruce, G., How to Click with your Horse; Clicker Training for Groundwork, Riding and Problem Solving, 2011, Port Douglas, QLD: Click with Horses.
8. McGreevy, P. and A. McLean, Equitation Science, 2010, Chichester, UK: Wiley Blackwell. 314.