You turn two horses out in a field together and watch with amusement as they interact. The horses, long-time friends, are thrilled to see each other; they squeal in delight, careen around in glee, roll ecstatically and vigorously scratch each other’s backs in greeting. Their joy is evident. Well, it certainly looks like joy…but is it?  The truth is, equine scientists know very little about horses’ emotions for the simple reason that horses cannot tell us what they are feeling. They cannot fill out questionnaires, participate in interviews or engage in any of the measurement techniques we use with human participants.

That said, there are some things about equine emotions that we know for sure (or that are agreed upon by most equine researchers), some things that we think we know (or that equine researchers are exploring), and some things that might be possible. I will discuss what we know about our horses’ emotional world, why we should make inferences in order to know more and why we should not.

What do we know for sure?

The short answer to this question is: not very much. Historically, animals were thought to be capable of only the most rudimentary mental processes. The idea of animals having thoughts or feelings remotely like our own threatened our uniqueness and subsequent superiority in the animal phylogeny. Studies of animal emotion were further thwarted by science’s credo to adhere to observable and measurable behaviours.

Although the study of animal emotion has grown since the advent of cognitive ethology (the study of animals’ mental processes) in the ’70s, the field is still in its infancy.

How do we know it?

Some researchers have attempted to measure equine emotions by asking their owners what their horses are feeling (Morris et al., 2008), but this is fraught with subjective interpretation. We do have some objective physiological measures (such as heart rate, blood analyses, cortisol levels and neurological imaging), many of which have been used successfully to measure equine anxiety. We may also be able to measure equine happiness or, more scientifically speaking, the activation of brain reward centres. For example, researchers have found an increase in opiate activity when rats are playing with each other.

Do horses experience fear? Since horses in their natural state experience high predatory pressure (i.e. they sit firmly on the “flight” end of the fight/flight continuum) they have been evolutionarily designed to experience fear and demonstrate it with a hair trigger flight response. Although, we have modified the horse’s environment such that wolf or lion attacks are unlikely, as we are well aware there are many sources of terror in our horses’ worlds. Furthermore, horses are also evolutionarily designed to have a very good memory for fearful experiences. And while all this may have served them well for living on the plains, it has less utility for them, and us, today.

Equine researchers, Lisa Leiner and Markus Fendt from the University of Tübingen in Germany, found that horses exhibit a consistent and predictable behavioural pattern in their escalation of fear to a novel object. Also, fear behaviours (snorting, avoidance, etc.) were strongly correlated with physiological measures of fear (heart rate), suggesting that with this emotion, at least, we can rely on our assumptions about horses’ behaviours. Horses that act afraid probably are afraid.

Leiner and Fendt also demonstrated that fear responses can be eliminated with habituation and desensitization training (gradual and incremental exposure to the feared object). Other research has shown that positive reinforcement training (using food to mark and reward desired behaviours) may be even more effective in that horses learn the desired behaviour more quickly, perform more readily, generalize more easily and remember the lessons longer (Sankey, 2010). It is clearly in our best interests to train our horses pre-emptively to tolerate more and more novel objects and situations in order to avert the extreme panic reactions from 1,200 lbs. of terror – a situation most of us would prefer to avoid.

Thus, the hard data tells us that horses most certainly experience fear and probably joy (or a least a physiological state of pleasurable arousal). After that, we stretch to making interpretations from observable behaviours and inferences from other animal studies.

What do we think we know?

Do horses love each other? Horses appear to have a complex social network, sophisticated social cognition (having a knowledge of self and others, and knowing, remembering and understanding the complexities of rank and relationships in the herd) and form specific attachments to other horses. Researchers have observed that horses living in natural circumstances engage in allogrooming (mutual and synchronous, biting around the withers and neck) with particular others and that these partnerships (dare I say “friendships?”) are enduring.

Allogrooming seems to have calming properties by lowering heart rate, decreasing cortisol levels, and increasing endorphin production. It is thought to cement pair bonds between horses and thus contribute to herd stability. Whether the reduced anxiety and pleasurable state is equivalent to the emotion of affection or love is yet to be determined.

Do horses love us? Andrew McLean and Paul McGreevy, prominent equine researchers from Sydney, Australia, are currently researching horses’ capacity to form attachments to their human caregivers. “Attachment” in this context refers to the theory developed by psychiatrist John Bowlby in the early ’60s to explain an infant’s tendency to seek proximity and form strong bonds with a caregiver who provided protection, comfort and support. Bowlby believed that this attachment system evolved over time and arose instinctively to increase the infant’s chance of survival.

The same calming properties of allogrooming can be mirrored in human-to-horse interactions with vigorous wither grooming. McLean suggests that this tactile contact may be comparable to the tactile comfort that an infant seeks with an attachment figure such as a parent or close caregiver, and thus forms the basis of an attachment relationship. Additionally, since in many modern equine management systems there is limited or no horse-to-horse tactile contact, this bond with a human owner may be particularly critical. Whether a horse could feel “attached” to a human in the sense that Bowlby conceived of it (using the attachment figure as a safe haven in times of stress and orienting to that figure for a sense of felt security), well that is fuel for another article! Certainly, if horses do “attach” to one another, then they would undoubtedly experience loneliness from separation and grief upon loss. Whether they have these same feelings toward their human handlers is still a question. If they do attach to their owners as they would to one another, sadly, this may be a consequence of having no equine alternative.

Do horses feel grief? Bowlby, who was influenced by Charles Darwin’s evolutionary theory, believed that grief, like attachment, was not a uniquely human trait and viewed the mainly anecdotal accounts of grief in animals as a piece of the wider evolutionary picture. The common reactions shown by human and non-human mammals in situations of loss and separation led Bowlby to view these processes as common biological reactions with similar underlying causes. Barbara King, author of the recently published How Animals Grieve, chronicles many accounts of animal grief – elephants who stay by the body of a deceased elephant for days, caressing and attending to their fallen herd mate, crows who bring offerings of sticks or bits of grass and lay them on or next to a dead crow and chimpanzee mothers who will not let go of an infant that has died even as the body is decomposing. Since many animals, including horses, are as intensely social as humans, it is reasonable to assume that they would experience similar feelings of grief upon loss or separation from close herd mates. King believes that the grief we see in animals is necessarily less complex than our own, as most animals lack the cognitive capacity for deep reflection and remembering. Nor would it be evolutionarily plausible for animals to indulge in the kind of grief that humans suffer even if they did feel it the same way; it is simply too energy costly to be dysfunctional with grief in a natural world where survival is precarious.

Are horses angry? Horses often show aggressive behaviours toward other horses and toward humans, but are they angry in the way that we can be angry? Some apparent anger or grumpiness in horses may be attributable to physical pain, and they may be in pain much more frequently than we think. For example, Carole Fureix and her associates from the University of Rennes in France found, not surprisingly, that horses who suffered chronic back pain were more aggressive toward humans than those who were pain-free. What is more surprising, was that in her sample of 59 school horses from three riding centres, with a light to moderate work load, 73 per cent were severely affected with vertebral pain. Note, that these horses were not recruited from veterinary clinics treating equine back conditions, but working school horses.

Dr. Sue McDonnell, a veterinarian and equine behaviour specialist at the University of Pennsylvania’s New Bolton Centre, says that 90 per cent of all their horses referred for behaviour problems (rearing, bucking, refusing to move out of the in-gate, self-mutilation behaviours, etc.) have a physiological cause – most commonly ulcers.

Also relevant is that horses, unlike humans, do not appear to hold a grudge. McDonnell adds that even with horses who have been in chronic pain for years, when the pain is resolved, so too are the behaviour problems, virtually overnight. Horses’ seemingly immediate forgiveness in this regard speaks to the unlikeliness of their experiencing the kind of anger that we do, that can translate to feelings of jealousy, hatred or revenge.

What emotions might be possible?

More complex emotions – including jealousy, revenge, pride, shame and guilt – what psychologists call “secondary emotions,” require higher order cognitive processes, selfawareness and an ability to take another’s perspective. In children, these emotions emerge toward the second year of life and beyond and are influenced heavily by familial and cultural values – value systems that are unlikely to offer an evolutionary advantage to horses.

Do horses have a moral code of ethics? It may be unnecessarily limiting to suggest that horses’ emotional world is devoid of complexity. Marc Bekoff, a cognitive ethologist, notes that it is not reasonable to assume that complex emotions evolved suddenly and uniquely in humans with no precursors in other animals. Rather, he argues, evolutionary continuity is more likely where differences in emotional capacities between human and non-human animals are differences in degree rather than differences in kind.

For example Bekoff studies dogs’ play behaviour and argues that they exhibit a rudimentary kind of morality that incorporates a set of rules of social engagement. Justice, cooperation, fairness and trust make the play work even when dogs are mismatched in size and strength. If a dog is inadvertently hurt during play, the offender will counter with a “play bow” as if to apologize and invite the play to start again. Bullies, who do not abide by the rules, are ostracized from group play, suggesting that dogs have a code of ethics that values fairness and cooperation. Since play is evolutionarily costly, there must be a payoff for its existence. Researchers have suggested that horses play to learn the roles and behaviours that they will later need to live cooperatively in a herd. But Bekoff goes further, arguing that play exists because it is fun and the joy associatedwith play may be so strong as to outweigh the risks of injury, and depletion of energy.

Bekoff maintains that we need to make interpretive leaps when trying to understand animal emotions and criticizes science’s reductionist perspective to that of observable behaviours. Instead, he argues, we need to explore animal emotions, to infer and to make educated guesses.

But are we able to make these inferences with any degree of accuracy? It appears that we do. In his groundbreaking research of the ’60s, Paul Ekman and his associates asked participants to describe what emotion was being expressed in a series of facial photographs. Hefound that people can reliably identify six basic emotions (happiness, sadness, anger, fear, disgust and surprise), and that these emotions appeared to be universal. When Ekman took his research to a remote area of Papau New Guinea, where participants had no prior Western culture exposure, they too identified the same six emotions.

Presumably, this ability might extend to reading animals’ emotional states which, especially in mammals, seem quite recognizable. Even people with little experience often agree with one another on what an animal is most likely feeling. Paul Morris found that humans’ attributions about animals’ (including horses’) emotional states predicted future behaviour quiteaccurately (2008). As Frans de Waal so aptly commented, “Sometimes I read about someone saying with great authority that animals have no intentions and no feelings, and I wonder, ‘Doesn’t this guy have a dog?’”

Why does it matter?

Bekoff makes a credible argument in saying that we need to stretch beyond what is directly observable in interpreting animal emotion, and there is evidence to suggest that we are reasonably skilled at doing so. However, I would counter that science’s reductionist principles keep us on track about what may or may not be going on inside an animal’s head.

Science has good reasons to be cautious about anthropomorphism (or attributing human thoughts, feelings and intentions to non-human animals). From the scientific perspective, our openness to numerous interpretations is compromised when we enclose that observable data into one fixed meaning. When we say that the reunited horses of this article’s opening are galloping joyously, we may miss a multitude of signals being exchanged between the two that might lead us down altogether different and productive research paths.

From the horse’s perspective, anthropomorphism can compromise his welfare when our interpretations of his behaviours are biased by our own goals and ambitions. Much of horses’ behavioural repertoire (both desirable and undesirable) is typically ascribed to their personality or emotional perspective. Horses’ exemplary behaviours are often attributed to their kindness, generosity or stoicism. Unwanted behaviours have been attributed to their malevolence, revenge or spite. The more likely explanation is that horses perform well because they have internalized a strong reinforcement history for doing so. Horses exhibit unwanted behaviours when they do not understand the question, have become desensitized to the aid by the continual application of simultaneous conflicting aids, endure physical discomfort when complying or are physically or psychologically incapable of performing the required behaviour.

When we take our interpretive leaps too far and believe that our horses perform well out of benevolence, we can also feel justified in punishing them when they purportedly perform poorly out of malevolence. And that punishment is rarely justified. It may be possible that horses are capable of complex secondary emotions, and perhaps future sophisticated brain imaging techniques will give us more insight here. In the meantime a more parsimonious explanation exists. Horses exhibit conflict behaviours or “misbehaviours” when there is disconnect between what is being demanded and their physical and psychological ability to comply. Indeed, equine researchers refer to these behaviours as “conflict behaviours” in an effort to remove value judgments about the horse’s moral fibre. We would do well to discover this conflict first before we leap to conclusions about our horses’ problematic attitude, poor work ethic or a personality turned bad.