Written by: Antonia J.Z. Henderson

Equine psychologist Antonia J. Z. Henderson, Ph.D., explores communication and understanding between humans and horses.

Thumbnail for Talking to Horses: Exploring Communication & Understanding

Lydia Keen Photo

Communication between species is fraught with complication and not just because horses can’t speak. Does it matter? Indeed, it can. Training and welfare are both impacted when we misinterpret equine communication or they misunderstand ours.

The concept of Umwelt, first coined by German biologist Jakob von Uexküll, from the German word meaning “surrounding world” or “subjective universe” refers to the species-specific perception of reality of an organism. Organisms create and shape their own Umwelt, guided by their unique evolutionary histories. Entrenched as we are within our own Umwelt, our ability to step into an equine reality is fraught with difficulty. When our horses do not perform as we would like we are quick to attribute this perceived shortcoming to a lack of intelligence, a poor attitude, or a malevolent nature, when the more likely interpretation is a mismatch of Umwelts.

These mismatched worlds become excruciatingly evident in our human-to-horse and horse-to-human communication. Indeed, our “subjective universes” are so different it is miraculous that we manage to communicate as effectively as we do. The miracle occurs because horses have ventured much further into our Umwelt than we have into theirs.

I will outline what makes our horse and human communication worlds so vastly different, explore some of the small inroads we have ventured into their Umwelt, and argue how horses have taken on the heavy lifting in bridging this divide.

Mismatch #1: The non-verbal world of “Equinese”

Language is so woven into our Umwelt that the word “communication” is often synonymous with “words” – language, speaking, reading, or writing to humans. Horses, on the other hand, evolved an effective visual system of communication, where vocalizations were rarely necessary. Horses living under natural conditions form stable, often life-long, social networks, where maintaining group cohesion is critical to the herd’s survival and continuity. Horses, like humans and other social species, have thus evolved a sophisticated level of “social cognition” – an understanding of each group member’s sex, age, relationship, personality characteristics and social rank. This information is generally communicated visually; most vocalizations occur on the rare occasions when herd members are out of sight.

There appears to be an evolutionary pattern where animals living in open grasslands (who can easily maintain visual contact) tend to rely more on visual, non-acoustic, communication, whereas forest-dwelling animals, such as the wild boar and its descendant the domestic pig, have a wide array of acoustic signals. Horses also have a more complex facial musculature than pigs do, allowing them to convey more information through facial gestures (Budiansky, 1996).

Mismatch #2: We say several things at once

Poorly versed as we are in the non-verbal world of the horse, our cues are often clumsy and lack the subtlety of horse-to-horse communication. As the classic image of the deadened school horse with a novice rider mercilessly thumping his sides or pulling on the reins would belie, horses have exquisite tactile sensitivity, responding to pressures too light for a human to feel. Using instruments for measuring human touch, researchers have found that horses’ sides, specifically the area that comes in contact with the human leg, are more sensitive that human fingertips (Saslow, 2010).

Rather, it is the delivery of conflicting aids that results in the apparent insensitivity of the dulled school horse. Bombarded with inadvertent, but contradictory push and pull aids, where no response alleviates the pressure, horses fail to learn which signals are meaningful. They eventually tune-out the noise, and silently endure their sentence. Horses who have been labeled “dead-sided” or “hard-mouthed” may have been denied the opportunity to react to a consistent, meaningful, and feather-light aid. They have likely entered a state of “Learned Helplessness” where an organism faced with chronic, inescapable, stress simply gives up trying, even when escape becomes available (Seligman, 1967).

Mismatch #3: We shout; horses whisper

To understand the horse’s communication finesse, we need look no farther than how they communicate with one another. “Ritualized behaviours” – the epitome of equine, non-verbal understatement and the basis of most horse-to-horse communication – are abbreviated versions of the actual gesture they represent. They carry the same meaning, but have evolved because they are less energy costly. Consider the clear message of a tail swish from a horse at the hay bale to a subordinate that was thinking of nosing in to have lunch. Horses that quickly learned the message from a swishing tail and avoided the painful kick that would inevitably follow were more likely to survive and pass on that behaviour to their offspring. Similarly, the horse that could maintain access to a resource without engaging in an energy-depleting battle was also more likely to survive and pass on this behaviour to future generations. And thus, a refined system of whisper body language evolved.

Horses that have been trained to carry this sensitivity over to their work under saddle may suffer from the unintentional signals or the heavy-handedness of a novice.

When our horses do not respond to our aids, we often start shouting. Bits become stronger, spurs longer and aids administered with increasing force. The recent arrival of the BellyBand ™ that protects the horse’s sides from spur rubs from an incessant, but likely ineffectual leg, is the equivalent to stuffing cotton in someone’s ears so as to protect their sensitive ear drums when we are shouting at them, rather than simply lowering one’s voice.


A hot topic in animal research is that of “referential communication,” which occurs when a sender directs the attention of a recipient through gestures – pointing, showing or touching – to a target and, in this way, informs the recipient of his or her wishes. If the gesture works, both sender and receiver focus their attention on the target. In humans, this ability shows up at nine to 12 months of age, when children will begin to use gestures such as pointing, showing or offering an object to a social partner (Bates et al., 1979). Referential gesturing is thought to be a rare form of communication, one that most likely evolved only in those species living in highly complex social societies such as humans, other primates, dogs, dolphins and ravens.

The most promising results have surfaced with dogs that seem to have a particular flare for correctly responding to human referential gestures about the location of hidden food. Ethologists Hare and Miklósi, who have conducted the most extensive research in this area, suggest that the domestication of dogs, which has fostered a unique interdependent bond with humans, has precipitated an enhanced ability for dogs to read human gestures, not seen in other species. Without prior learning, dogs can understand and act upon a wide variety of human gestures from the first trial, and generalize this knowledge to new situations and new gestures.

Horses too form complex social societies, and share with dogs this very particular, interdependent relationship with humans. Rather less convenient research subjects than dogs, referential communication research is relatively limited in human-to-horse interactions. To date, the jury is out. There is some evidence that horses can follow human pointing gestures when the cue is close to the target (i.e. a bucket containing a carrot or oats jack-pot), but it is unclear whether they really understand what the gesture means. In one experiment, many horses nuzzled and followed the experimenter’s extended arm down the fingers to a reward bucket (Maros et al., 2010), suggesting that the horses’ success may have had more to do with the knowledge that human hands bring treats, than demonstrating a true understanding of human intention.

British researcher Leanne Proops and her group tested this ability in 28 horses to see whether horses could respond correctly when cued to find a target reward bucket. Horses seem to be able to make use of some human gestures (pointing cues, and placing a colourful marker in front of the target bucket) but not others (tapping on the bucket, orienting one’s body toward the target bucket or gazing at the target bucket).

Yet, other research suggests that horses may indeed be able to grasp human pointing gestures, but they need practice. In a study exploring the impact of previous training method (Parelli horsemanship vs. traditional methods) on horses’ ability to follow human gestures, Nicole Dorey and colleagues found that all horses regardless of their background initially performed dismally. However, with practice, all horses showed remarkable improvement. Interestingly, 90 per cent of the Parelli trained horses (which relies on human movements and gestures to guide a horse’s behaviour from the ground) reliably followed the researcher’s gestures, whereas only 40 per cent of the traditionally trained horses did so. Dogs show a similar effect; those with a prior history of relying on human gestures (as in agility trained dogs) outperform dogs without this background (Marshall-Pescini et al., 2009).

While horses may still be in the baby stages of following our pointing cues, when they need to alert us to something that they want, they show striking skill at implementing their own repertoire of referential gestures. Italian researcher Rachele Malavasi and Ludwig Huber, from Austria, tested 14 horses’ ability for referential communication with a human handler to obtain a visible, but out of reach food reward. Malvasi and Huber were interested in whether horses could understand differences in the attention of the handler (either toward the horse or elsewhere) and modify their approach accordingly. In contrast to gestures that are tuned in to the mere presence or absence of an audience, referential gestures take into account the receiver’s attentiveness. If, for example, a dog is choosing between whining or gesturing with his head to get a human handler to give him a desired object, he will only choose the latter if his audience is already looking at him. If this is not so, he will first engage in some attention getting behaviours such as whining, pawing, running about, etc.

These researchers found that horses used sophisticated and creative referential gestures to engage the handler’s attention toward the desired reward bucket, including alternating their gaze between the human handler and the bucket, nodding their heads, turning their tails and whipping their heads around quickly as if pointing to the reward bucket. In true “referential gaze” form, horses only made these efforts if the human handler was looking at them. If the handler’s orientation was elsewhere, they first tried various strategies to gain her attention – some even walking back to the researcher and pushing her toward the bucket.


We have had high expectations for horses to understand our Umwelt, while making rather pitiful advances into theirs. Researcher Steven North argues that our anthropocentric, or human-centered, bias hampers our understanding of the impact of our interventions in a multispecies world. While North focuses on horses’ interaction with technology, he contends that his premise extends to all human/horse interactions. Our methods of dealing with horses have become so commonplace, and horses’ behavioural responses sufficiently subtle, that we may well have become desensitized to less overt signs of distress.

North notes that horses are “unaware interactors” – i.e. the horse is unaware of the underlying system, the intention and the purpose behind most of our human interventions. This places a greater responsibility on us to ensure that the animal’s welfare is guaranteed. North says, “unaware interactors are entitled to interactions that enrich rather than exploit.”

The horse has come a long way into our world. They use their athleticism and their intellect to configure themselves in whatever version of a horse we have invented for them in each equine discipline. They do so with stoicism and with remarkably little protest. Perhaps this is so because a horse, for the most part, is the quintessential “nice-guy.” Unless we have done something to shatter their trust, most horses are affable, imminently social and extraordinarily generous in their efforts to cope with the oft-times taxing environment we have created for them. These highly social beings do their damnedest to communicate with us. Perhaps we should do more to venture a little further into their Umwelt and learn a little equine communication on the journey.


Researchers are making more advances in trying to understand the equine communication Umwelt through technologies that seek to identify and record a full repertoire of horses’ body language in various contexts. The “Equine Facial Action Coding System (EquiFACS)”, developed by Jen Wathan and colleagues is a comprehensive manual with accompanying photographs and videos that has identified, photographed and described all the possible facial movements of the horse. Wathan’s group has created an anatomically based, objective system, describing equine facial expressions across a range of contexts, reflecting different emotional states. The creators comment that horses have a rich facial repertoire (slightly fewer expressions than humans, but more than chimpanzees), and that many of the facial movements seen in horses are similar to movements seen in other animals – including humans!

They note that even common and recognizable facial expressions may have different meanings in different contexts. The EquiFACS ensures that observers focus on the individual facial movements, without preconceived notions about the expression’s meaning. Valuable information is retained by not leaping to what we think we “know” to be true. For example, the flehmen response (the upper lip curl typical of stallions when smelling good things like manure and mares in heat) has also been observed as an expression of pain. This latter interpretation could easily be missed without studying all the facial nuances that accompany the lip curl and the context in which it is occurring.

The EquiFACS provides a standardized language for sharing information among horse industry professionals and researchers. By providing the full capacity of horses’ facial expressions it has the potential to extend our knowledge of equine communication and cognition and facilitate research that can enhance equine welfare (Wathan et al., 2015). See tinyurl.com/EquiFACS to learn more.

Another such endeavour into the equine Umwelt is recent work by researchers Steve North, Many Roshier, and Carol Hall who are currently creating the “Horse Automated Behavior Identification Tool (HABIT)” – a computer program that automates the analysis and recognition of horse-to-horse, and horse-to-human behaviours, from video footage. The program assesses horses’ non-verbal behavioural responses in everyday interactions such as handling, training, housing, and feeding. North says that his group hopes to answer questions about equine welfare such as how horses are managing with the environmental constraints that we have created, and if they are still able to live within the normal behavioural range for their species despite these environmental limitations. Since we cannot interview horses and ask them how they are feeling, systems that can detect nonverbal behavioral responses to interventions can be enormously powerful in ensuring equine welfare.