Yes, but “recognize” needs a precise definition. Controlled experiments show that horses respond differently to selected positive and negative human facial expressions, nonverbal vocalizations and combinations of faces and voices. One study also found that a horse’s later response to a person could differ according to that person’s expression shown hours earlier.
These results support discrimination and association of tested emotional cues. They do not show that horses consistently identify a person’s private feelings, understand complex human emotions as people do, or read minds during every everyday interaction.
As an Amazon Associate I earn from qualifying purchases.
What scientists mean by “recognize human emotions”
In ordinary conversation, recognizing an emotion might mean knowing that someone is angry, frightened or sad. Animal-behavior experiments use a narrower standard: researchers present a defined cue and measure whether an animal responds differently to one version than to another.
Recommended Free Tools
For horses, the measured cue may be a photograph of a smiling or angry face, a recording of laughter or growling, or a face and voice presented together. The measured response may be gaze direction, heart-rate change, freezing, vigilance or behavior toward a person shown later. A difference between conditions demonstrates that the horse detected and used information in the stimulus. It does not, by itself, reveal exactly what the horse thought the person was feeling.
#1 Best Overall
This distinction matters because a face, voice, posture, movement, smell and previous experience can all occur together in real life. The experiments isolated particular inputs so the horses’ responses could be compared fairly.
What the main experiments found
| Study and year | Human cue | Person familiarity | Measured response | What the result supports |
|---|---|---|---|---|
| Smith and colleagues, 2016 | Photographs of positive (happy) and negative (angry) facial expressions | Tested photographs; the summary does not establish that the people were familiar to every horse | Angry faces produced a left-gaze bias and a quicker heart-rate increase | Different responses to the tested facial stimuli |
| Smith and colleagues, 2018 | Positive and negative nonverbal vocalizations: laughter and growling | Playback recordings; no other emotional cue was presented | Greater vigilance, including freezing, to negative vocalizations | Horses distinguish at least some vocal emotional cues |
| Trösch and colleagues, 2018, Scientific Reports | A human face paired with a voice carrying positive or negative emotional value | Responses included presentations involving unfamiliar people | Behavior reflected the relationship between the visual and auditory cues | Evidence consistent with cross-modal association of emotional valence |
| Proops and colleagues, 2018, Current Biology | A photograph of a specific person displaying an angry or happy expression, followed hours later by that person in a neutral state | Person-specific test; the later person matched the earlier photograph | Responses to the previously seen person differed from responses to a mismatched person; human observers were blind to which photograph had been shown | A lasting, person-linked association with the earlier expression |
The table separates four kinds of evidence that are sometimes collapsed into the single phrase “horses read emotions.” Immediate facial discrimination, response to an isolated voice, integration of face and voice, and later person-specific memory are related but not interchangeable abilities.
Facial expressions: horses reacted differently to happy and angry faces
In the 2016 experiment by Smith and colleagues, horses viewed photographs of humans displaying positive, happy expressions and negative, angry expressions. The reported differences were specific: angry faces were associated with a left-gaze bias and a quicker increase in heart rate.
PC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchA left-gaze bias is an observable direction-of-gaze pattern. It should be reported as such rather than translated into an unproven statement such as “the horse knew the person was angry.” The quicker heart-rate increase is a physiological response showing that the two facial conditions were not treated identically. Neither measurement identifies the horse’s exact subjective experience.
Rank #2
The result is still important. The photographs supplied visual information without requiring the horse to hear the person’s voice or watch a live body. Under those controlled conditions, the horses responded differently to the positive and negative facial stimuli.
Voices: negative vocalizations increased vigilance
Smith and colleagues’ 2018 playback study tested human vocal emotion without adding a face. The recordings were positive and negative nonverbal sounds, described in the study summary as laughter and growling.
Horses showed increased vigilance to the negative vocalizations rather than the positive ones. Freezing was among the reported behaviors. Because the study used audio playback, the difference cannot be attributed to a simultaneously visible facial expression. It demonstrates sensitivity to information carried by the sound itself, at least for the vocalizations selected by the researchers.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
“Negative” here describes the emotional value assigned to the experimental recordings, not a claim that every growl will have the same meaning to every horse. A horse’s history, the loudness and context of a sound, and the identity of the speaker could all matter in situations outside the test.
Face and voice together: evidence of cross-modal association
The 2018 Scientific Reports study by Trösch and colleagues asked whether horses could relate what they saw to what they heard. Horses were shown a human facial expression while hearing a voice with positive or negative emotional value.
The authors reported evidence consistent with cross-modal perception: the horses’ responses reflected the relationship between the face and the voice rather than treating each channel as entirely separate. The experiments also included unfamiliar people, and the reported effect generalized beyond a single known individual.
This is stronger evidence than a simple reaction to one isolated picture or sound because it tests whether information from two senses is associated. Even so, “cross-modal” does not mean the horse inferred a complete private mental state. The tested conclusion is narrower: visual and auditory cues with emotional value could be connected in the horse’s response.
Memory: an earlier expression influenced a later response to the person
Proops and colleagues’ 2018 Current Biology study examined whether an expression could remain associated with a particular individual. First, horses saw a photograph of a specific person displaying either an angry or a happy face. Several hours later, they encountered that person in a neutral state.
The abstract reports that the horses’ responses to the previously seen person differed from their responses to a mismatched person. The human observers were blind to which photograph each horse had viewed, reducing the chance that an observer’s expectations caused the behavioral difference.
This design tests more than an immediate startle or gaze response. It indicates that the horse could retain an association linking an individual with the earlier facial cue and use that information later, when the person’s face was neutral. The result does not establish how long such an association lasts beyond the tested interval, nor does it show that the horse remembered a human emotion in the same conceptual way a person would.
What the evidence supports—and what it does not
Supported by the experiments
- Horses can show different behavioral or physiological responses to selected positive and negative human facial stimuli.
- They can respond differently to selected positive and negative nonverbal vocalizations presented without other emotional cues.
- They can associate emotional information from a face and a voice, including in tests involving unfamiliar people.
- They can retain a person-specific association with an earlier facial expression for at least the several-hour interval tested in the memory experiment.
Not established by these results
- That every horse can reliably tell when a person is sad, angry, afraid or happy in ordinary life.
- That horses understand complex human emotions, intentions or private thoughts.
- That a physiological or vigilance response has one unambiguous emotional meaning.
- That facial cues are always more important than voice, posture, movement, prior learning or other context.
- That the findings justify a population-wide percentage or a universal rule for all horses and people.
Why “cue discrimination” is more accurate than “mind-reading”
A horse may learn that a particular face, voice pattern or person is associated with a positive or negative outcome. It may also be sensitive to changes in pitch, facial configuration or other features that humans label as emotional. Those abilities can produce an appropriate-looking response without requiring a human-like concept of anger or happiness.
Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Repair Windows errors before they cause bigger problems3Scan for outdated or missing drivers - takes under a minuteResearchers therefore describe the results in terms such as discrimination, association, valence and cross-modal perception. These terms keep the claim tied to what was measured. Saying that a horse “read the rider’s feelings” goes beyond the evidence unless the situation has been tested and the relevant cues are known.
Best Value
How to use this knowledge around a horse
The studies offer a useful reason to pay attention to your signals, but they are not a diagnostic tool for deciding what a horse “knows” about you. A practical approach is to treat emotional communication as multi-channel and observe the horse’s actual behavior.
- Notice your voice and face. Horses may respond differently to vocal and facial cues, so avoid assuming that changing only one signal will leave the overall message unchanged.
- Watch for the horse’s observable response. Changes in gaze, heart-rate-related arousal, vigilance or freezing are responses to monitor, not proof of a specific human emotion being understood.
- Consider the whole context. The experiments isolated cues; real interactions also include posture, movement, distance, environment and the horse’s prior learning.
- Do not force an interpretation. If a horse reacts, describe what happened—such as increased vigilance or a change in gaze—before assigning a label such as fear, anger or empathy.
- Use calm, consistent handling. When a horse is alert or frozen, give it space and time rather than escalating the stimulus. The studies do not provide a training protocol, so ordinary safety and qualified professional advice remain important.
Important limits of the current findings
The evidence comes from controlled presentations of selected photographs and recordings, with responses measured under experimental conditions. That design is valuable because it isolates variables, but it cannot reproduce every feature of a stable, trail ride or training session.
The published findings also do not settle how horses weigh facial expression against body posture, movement, smell, voice, familiarity or past experience in daily life. The cross-modal work shows that visual and auditory information can be associated; it does not rank one sense as the horse’s dominant source of information in all circumstances.
Nor do the studies establish human-like empathy. A horse’s different response to an angry face or a growl may reflect learned associations, sensitivity to perceptual features, general arousal or several processes at once. The experiments demonstrate a difference in response, while the underlying subjective experience remains open.
Bottom line
Horses can distinguish and associate certain human emotional cues. Experiments have found different reactions to happy versus angry faces, greater vigilance to negative versus positive vocalizations, links between matching face-and-voice signals, and a person-specific effect that persisted for several hours. The careful conclusion is not that horses invariably know how humans feel, but that they are capable of detecting and using selected emotional information in ways that can influence their behavior and physiology.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
Leave a Reply