5 Points to Consider in Horse-Human Attachment Research

Can horses attach to humans?

It is certainly a valid question. Arrazola and Merkies (July 2020) reviewed the literature on horse-human attachment research and found studies showing horses responding to humans as safe havens and secure bases, and showing proximity seeking and separation distress in relation to people as well. However, additional research is required, as studies continue to get published that question whether horses are capable of attaching to humans in the same ways as other people and even dogs do.

Mainstream media write-ups of such research don’t help make sense of the situation; titles such as “Horses DON’T form attachment bonds with their owners” (November 2020) are based on the results of one study and don’t take the bigger picture into consideration. There is far more nuance at play than sensational titles propose. The research is ongoing and I encourage people not to jump to the same conclusions as these titles do.

With that in mind, I read the following article by Hartmann et al., “From the horse’s perspective: Investigating attachment behaviour and the effect of training method on fear reactions and ease of handling – A pilot study“, back in February 2021 and began this draft, but did not complete it until now. There are a few points I think are worth drawing attention to, that I warmly invite researchers to consider in conducting future research in this intriguing area of study.

1. The flight animal myth

This is a small yet important point: The research article calls horses “flight animals”, which perpetuates a horsemanship myth. Other prey species also flee but are not given such a label; De Giorgio (co-author of Equus Lost?) says “We humans create what we think is a flight animal by trying to control what we think is a flight animal”. Horses are also capable of other survival responses, yet they are not called “fight animals” or “appeasement animals”, for instance. This narrative comes out of the horsemanship industry and creates a justification for using certain methods. They are prey animals, certainly. But flight animals is a term that would be wise to lay to rest given its connotations. See his video here on “The Flight Animal Fallacy“.

2. Attachment requires more than familiarity

The study found partial support for the idea that humans can be attachment figures for horses, but this study like other similar ones worked off the premise that attachment derives from familiarity. That is, that horses should show an attachment relationship with their owners because they are familiar with them. However, attachment theory is also clear that familiarity is not enough. If familiarity was all that was required to become attached, then we would show secure attachment behaviours like safe haven, secure base, proximity seeking and separation distress towards other familiar people in our lives regardless of their role (e.g., bus driver, massage therapist, etc.). This line of thinking also presumes that a child (or horse) would show secure attachment behaviour towards a parent or caregiver (or owner) even if said adult did not provide the conditions for secure attachment.

A woman with a horse

We become securely attached when we are seen, heard, feel felt, and get gotten. Attachment theory posits that attachment is about these kinds of relational features, and not about the presence of a reinforcer like food. How many adult humans do you know that were neglected, mistreated emotionally, or felt misunderstood and struggle in relationships yet express guilt because their caregivers provided them with food, a roof over their heads, and other physical necessities of life? The presence of the latter does not mean that all is well with the former.

Early behaviourists thought attachment occurred between an infant and their caregiver because of being provided nourishment and that attachment was learned. However, attachment research decades ago by John Bowlby, Mary Ainsworth, Judith Solomon, Mary Main, and others since then has demonstrated that mammals are born with an inherent drive to attach and that secure attachment is developed through nurturance and responsiveness to needs and cues. Are relational attunement and repair missing in the horse-human pairs in the studies that have looked at attachment so far?

At least one study of horse-human bonding looked primarily at the role of food in creating attachment bonds (Sankey et al., 2010), but the presence of treats may generate an interest and proximity seeking because the treats may be novel as opposed to evidence of a “bond”. Even this most recent study by Hartmann et al. (2021) sought to explore attachment through behavioural principles; more specifically, they looked at the possible role of positive reinforcement rewards via treats and withers scratches (versus negative reinforcement) in forming attachment bonds with handlers. I have written elsewhere that connection must come before concepts are applied (or, as Warwick Schiller now says, “relationship before horsemanship”). Will providing treats and scratches to an animal that may have an already underlying insecure attachment style with humans due to past experiences (known or not) lead them to develop secure attachment? It is questionable whether external reinforcers will create deep, wholesale change in an animal’s existing relational neurophysiology beyond temporary overriding, without also providing the relational conditions that support secure attachment. Same with humans – providing snacks while remaining misattuned is not likely to produce secure attachment, no matter how tasty the morsels are. This is akin to believing that star charts and praise will lead a child to become more securely attached to an emotionally absent caregiver. Are these and other studies putting the proverbial cart before the horse? This is not to say that the humans in these studies are emotionally absent or misattuned, of course — only that these factors are not being assessed.

Attachment is also, at least in part, determined by personality traits, which may also be mitigated by past adverse experiences.

The difficulty, of course, lies in the fact that the term attachment is used loosely in the research literature; some studies refer to attachment through the lens of attachment theory, others mean it more generally in terms of a “bond” more generally-speaking, and others use a mix (attachment theory language with a behavioural lens). This adds to the confusion and may also explain the conflicting results. Certainly, humans do bond over food, but this is a primate trait – equines graze together but do not have the same cultural layers associated with the social ritual of eating as humans do (Kieson & Abramson, 2018). And, of course, the conditions of secure attachment need to be present for there to be secure attachment – otherwise, bonding over food is merely affiliative and not evidence of anything deeper.

If using attachment theory to look at attachment dynamics in horse-human relationships, then I would like to see a study that looks at secure vs insecure attachment patterns and the characteristics of the owners that may contribute to the dynamic playing out (such as nervous system dysregulation, mental health challenges, misattunements, confusing or inconsistent reactions, punitive behaviour, gaslighting, being disconnected from oneself and others, etc.). This is not to shame or cast blame, but rather to objectively take full account of what may be playing out.

It would also be interesting to look at the humans’ trauma and attachment histories to see if there is any correlation there as well. To my knowledge, there have been no studies that evaluate both the horse and the human’s respective attachment styles in evaluating specific pairs’ dynamics. Arrazola and Merkies (2020) did measure the humans’ attachment styles and looked at equine behaviour towards these humans as either affiliative, avoidant, or oral, though correlating these 3 categories of behaviour with attachment styles in horses has yet to be done. It would also be interesting to add in a measure of the horse and the human’s personality traits as another mediating factor involved in attachment. Could attachment behaviours be missing or inconsistent due to a mismatch in personality traits or type between horse and human (which would affect the ability for attuned responsiveness)?

3. Potential diagnostic overshadowing

Of course, the field of attachment research would benefit from a study that considers the horses’ backgrounds in greater depth, comparing the attachment responses of horses who have a developmental trauma history to those that do not have such a history towards a human they are in regular contact with. For example, do horses that seem to show no or partial attachment behaviours have a history of imprint training, early weaning, and other early attachment disruptions (not to mention other traumas at the hands of humans at other points in development and adulthood), which can lead to wariness and insecure attachment towards humans (Henry et al., 2006; Henry et al., 2009)? Is what we are seeing evidence of anxious, avoidant, or disorganized attachment as opposed to secure attachment?

So many domesticated horses have experienced early traumas that is is ubiquitous. This means that we risk forming opinions about horse behaviour that are based on how they cope under problematic captivity conditions (as opposed to how they actually are); this is the point of ethology as an area of study, as championed by Lucy Rees and others. This begs the question: are the research results that keep coming up actually an example of diagnostic overshadowing? As in, we see a lack of attachment behaviours in horses and assume they are not capable of it, when in fact this may simply be a reflection of early attachment ruptures and/or trauma (and/or current relationship experiences with humans) and not evidence of their true capacity? This has been found in different human populations. For instance, autistic individuals are assumed to lack empathy when in fact their experiences of trauma and overwhelm may actually drain their capacity for empathy (Hume & Burgess, 2020), which speaks volumes about society’s ability to produce non-traumatized autistic people (as highlighted by Neurodivergent Rebel).

If this is the case, the body of research that questions whether or not horses are capable of attachment to humans may actually be pointing to a different issue: what are the pervasive captivity conditions that are contributing to insecure attachment in horse-human dynamics and what can be done to change those? To paraphrase Neurodivergent Rebel, is it possible that society is not doing a good job of producing non-traumatized horses that have difficulty attaching to people? What needs to be different?

4. Neuroception of safety… or not

Perhaps a study that is informed by the polyvagal theory would be enlightening. For instance, what would neuroception look like in horse and owner pairs?  Do horses neurocept safety, danger, or life threat in the presence of “their” human? What does “their” human neurocept in relation to the horse? According to Stephen Porges, attachment and connection are not possible if we do not have a felt sense of safety in relationship (2018). Tracking HRV, calming signals, and stress responses in both horse and human as the human approaches their horse from a distance and gets closer would also be interesting. Most studies on horse-human attachment (like studies on dog-human attachment) use an adaptation of the Strange Situation Procedure. However, that protocol involves the horse and human starting off together so there is no baseline of horse and human apart before beginning. Activation (anxiety, fear) or detachment (avoidance) about attachment shows up way before coming together; backing things up to an earlier point in the sequence and examining what is going on as horse and human approach one another may yield interesting information that might give a better sense of the nuances than simply evaluating the presence or absence of secure attachment behaviour.

It would also be interesting to track HRV in both horse and human pairs to see if they sync up or if they show evidence of co-regulation and social coherence in Strange Situation testing scenarios.

5. Adapting criteria based on population characteristics

Post-traumatic stress disorder (or PTSD) in humans has a specific set of diagnostic criteria. These criteria are different in children, for whom the signs of PTSD manifest differently than in adults (Rittmansberger et al., 2019; Wigham, Hatton & Taylor, 2011). Within the literature, there is also a movement that is advocating for adapted criteria for autistic people and individuals with learning disabilities, who manifest signs of PTSD differently as well. In a world that privileges neurotypical presentations, accounting for neurodivergence is important.

Similarly, it may be worth considering adapting attachment classifications based on species, especially based on brain structure and functioning (“neurodivergence” between species). Attachment theory research in dog-human pairs may have different outcomes than horse-human pairs simply by virtue of the fact that dogs and horses have different brain structures. This would mean that it is not that horses cannot attach to humans, but rather that how attachment looks in horses may very well be different and that the criteria for attachment may need to vary between species.

That said, others have pointed out that domesticated dogs share our lives and homes in ways that horses do not, and that this may also account for differences in attachment patterning between the two (Payne et al., 2016). The point still remains, however: is it that horses cannot attach, or that there are other external variables that affect the capacity for attachment?

Of course, it is entirely possible that horses are not capable of attaching at all – a possibility that seems unlikely to the most passionate proponents of the horse-human bond and that flies in the face of Arrazola and Merkies’ literature review (2020). However, before jumping to this conclusion, it would be wise to accurately define the terms and study these underlying factors that may explain the mixed research results before fully ruling out the possibility that they can. There is still a long way to go, and I applaud all research that attempts to shed light on this area of inquiry.


Arrazola, A. & Merkies, K. (2020). Effect of human attachment style on horse behaviour and physiology during equine-assisted activities: A pilot study. Animals, 10(7), 11156.

Henry, S., Richard-Yris, M.-A., Hausberger, M. (2006). Influence of various early human–foal interferences on subsequent human–foal relationship. Developmental Psychobiology, 48, 712–718.

Henry, S., Richard-Yris, M.-A., Tordjman, S., & Hausberger, M. (2009). Neonatal handling affects durably bonding and social development.  PLOS ONE, 4, e5216.

Hume, R. & Burgess, H. (2020). “I’m human after all”: Autism, trauma, and affective empathy. Autism in Adulthood. Published online ahead of print: https://doi.org/10.1089/aut.2020.0013

Kieson, E. & Abramson, C.I. (2018). Horses are not people: Research, parallels, and problems using equines as relationship models in EAP. In I. Parent (Ed.), A Horse is A Horse, Of Course: Compendium for the 2nd International Symposium for Equine Welfare and Wellness. Createspace Independent Publishing Platform.

Payne, E., DeAraugo, J., Bennett, P., & McGreevy, P. (2016). Exploring the existence and potential underpinnings of dog-human and horse-human attachment bonds. Behavioural Processes, 125, 114-121.

Porges, S.W. (2018). Trauma and Intimacy Through the Lens of the Polyvagal Theory: Understanding the Transformative Power of Feeling Safe. United States Association for Body Psychotherapy conference seminar. Santa Barbara, California (November 3).

Rittmannsberger, D., Kocman, A., Weber, G., & Lueger-Schuster, B. (2019). Trauma exposure and post-traumatic stress disorder in people with intellectual disabilities: A Delphi expert rating. Journal of Applied Research in Intellectual Disabilities, 32(3), 558-567.

Sankey, C., Henry, S., Gorecka-Bruzda, A., Richard-Yris, M.-A., & Hausberger, M. (2010, November 15). The way to a man’s heart is through his stomach: What about horses? PLOS ONE, 5(11), 1-4.

Wigham, S., Hatton, C. & Taylor, J.L. (2011). The effects of traumatizing life events on people with intellectual disabilities: A systematic review. Journal of Mental Health Research in Intellectual Disabilities, 4(11), 19-39.