Andrew McLean – Cavallo magazine, July, 2005

Translation from German VOICE – Cavallo magazine, July, 2005

Horses have remarkably good hearing – apart from having large and independently mobile ears, their audible frequency ranges far above and far below that of humans. So there is no doubt that the horse can hear the human voice. Whether it always responds to it is another matter. Human beings attempt to interact with horses using their voices in three ways: as soothing tools, as voice cues and as rewards. Each of these can be effective, but the effectiveness of each varies because of the greater training skill required with regard to the use of the voice as a cue and as a reward. Most trainers are highly ineffective in these areas because they do not optimise the use of what is known as ‘learning theory’ to implement the use of voice in training.

There is plenty of anecdotal evidence among horsefolk that the dulcet tones of the human voice can be soothing to horses in lowering anxiety. These calming tones may be analogous to the similar tones used by mares and foals and other equines bonded in kinship. Perhaps too the presence of quiet, soft sounds means the absence of aggression which is also comforting. Many horsepeople claim their horses can recognise their voices amongst other human voices, and considering the hearing abilities of equines, this is very likely to be so. However the horse is not born with the human voice in its head. It is only for the last 5 or 6 thousand years that we have domesticated them. For many more thousands of years before that, we chased, ambushed, trapped and killed them. If anything, the sound of the human voice should be innately terrifying to horses. In fact loud, sharp, erratic and piercing noises are frightening for most horses.

Horses are essentially silent animals. Like most prey animals, their evolution partly depended on staying quiet to avoid detection by predators. The only real exceptions to this rule in the animal kingdom are those animals that are camouflaged (frogs, insects etc) and those that can sleep far out of reach (monkeys and birds). It makes little sense to be constantly making a din and thus advertising your presence, so horses evolved to be quiet communicators. We on the other hand evolved to be quite the opposite; language is our most important communication tool. Our primate ancestry bequeathed to us a tendency to chatter away almost continually, and we are masters of noise. Our words symbolise every possible event and feature that occurs on our planet. Talking is quintessentially human.

Dressage is, when executed correctly, remarkably correct in terms of animal cognition and learning theory. However things have changed since the horse was once an invaluable commodity for war, transport and agriculture. In those days, no horse could be spared and training methodologies had to evolve teaching programmes that allowed no horse to ‘slip through the net’ through not making sense of the training regime. Nowadays however horses are an expendable commodity, largely because they are used for sport. Using horses for sport means that if the horse can’t cope, then there will always be another to take its place.

As a sport, dressage is silent – it is not permissible to talk to your horse during a dressage test and the origins of this practise can be seen as correct in terms of learning theory. Studies on the acquisition of cues show that animals can easily learn multiple cues for a single response. So a horse, for example, can easily learn to respond to leg aids, seat aids and voice aids for the same response. After all, the environment is replete with various cues that predict particular events and animals have evolved to recognise them. However these studies also show that there is a limiting factor here too – if you train a particular number of cues then you need to use all of them each time or else you risk a lowering in responding by the animal. Humans are not particularly good at multi-tasking, so it seems that it would be better to limit the number of cues and stick to them consistently. Thus, with the array of leg, rein and seat aids, it is the voice, if any, that should be discarded. Research has shown that responding can also lower when multiple signals arise from different sense modalities. For example, dressage implicitly involves cues based on touch so there may be the chance of lowered responding with the addition of voice as a further cue.

Despite the general belief by horsefolk, voice is ineffective as a reward, except under very precise circumstances. Voice rewards are not innately recognised as such by horses because they are not primary reinforcers like food or freedom or petting. While the value of food as a

reward is not disputed, there are some impracticalities surrounding its use in horse training. Nevertheless the use of ‘secondary reinforcement’ where a sound signal precedes food in horse training (for example clicker training) has shown great results when practised correctly. Secondary reinforcement can and should be extended to other areas of training, and not just with the use of food. Many studies also show that petting the horse at the base of the withers relaxes the horse – it lowers heart rate and blood pressure, and therefore, can also be considered as primarily rewarding. Horses are commonly seen raking each other there with their teeth which strengthens familial bonds. This site at the base of the withers is also very conveniently located for the rider: just below the hands! Therefore, the most effective use of voice is to use it in the format of secondary reinforcement, like clicker training, where the term ‘good boy’ marks the correct behaviour and is followed by wither petting immediately or up to two seconds later.

To train the use of that voice cue, it is essential to use it precisely when the horse is giving the correct response and then always follow it up with petting at the base of the withers. The voice command can be thought of as the promise of petting and the trainer should not break that promise. The petting itself can be without the voice command but not the other way around. It is also important to maintain consistent sonic characteristics of the voice cue, such as word, tone, pitch and duration for effective use. In the early stages of acquisition of a learned response, the voice-reward sequence should occur for every correct response (continuous reinforcement schedule). When the desired behaviour becomes consolidated then the voice-reward sequence should occur occasionally (variable schedule of reinforcement) as this makes the animal work harder for the reward which is now rationed.

However, in the early stages of horse training where stronger leg or rein pressure is used, the reward is the release of pressure and no other rewards should be used. In fact at any point of the horse’s training where pressure is used (including anything that involves negative reinforcement such as the dressage whip), the trainer is most effective if he stays silent and refrains from voice commands, because studies show it is less effective to combine positive with negative reinforcement simultaneously. It is when the response becomes elicited by light aids that positive reinforcement (the voice-reward sequence) is best utilised.

So when we attempt to communicate with horses, we have to remember that we are the ones who chatter away, and maybe horses don’t need to hear it. The more we chatter away the greater the chances of dulling our signals to horses, because of the endless stream of background noise that may surround an occasional voice cue. Horses soon switch off! However, we must never punish the horse for what we expect it to understand or for how we expect it to respond amongst a barrage of background ‘white-noise’. Because the horse’s mental abilities are different from humans, we need to recognise that we have to be very clear in our training – that the horse is not able to extrapolate or fill in the gaps of our blurry attempts at communicating. Horses are sentient only for each passing moment of time, and have no concept of time itself. Like a rock in an ever-rolling stream, the horse is etched with all of the events in its life, the good and the bad, the clear and the unclear. But the horse cannot recall these events at will – only similar situations will arouse the memories locked away and frozen in time. To the horse the past and the future are entirely unknown and of no interest. It is up to us as, custodians of these extraordinary animals, to train them as carefully and correctly as human knowledge makes possible.