IN-DEPTH: BEHAVIOR

The Thinking Horse: Cognition and Perception

Reviewed

Evelyn B. Hanggi, MS, PhD

Cognition and perception in horses has often been misunderstood. Not only in the past but even
today, people proclaim that horses react only by instinct, that they are just conditioned-response
animals, that they lack advanced cognitive ability, and that they have poor visual capabilities (e.g.,
acuity, color vision, depth perception). Until relatively recently, there was little scientific evidence
to address such beliefs. Change, however, is underway as scientific and public interest in all aspects
of equine learning and perception intensifies. A review of the scientific literature, as well as practical
experience, shows that horses excel at simpler forms of learning such as classical and operant
conditioning, which is not surprising considering their trainability when these principles and prac-
tices are applied. Furthermore, horses have shown ease in stimulus generalization and discrimina-
tion learning. Most recently and unexpected by many, horses have solved advanced cognitive
challenges involving categorization learning and some degree of concept formation. A comprehen-
sive understanding of the cognitive and perceptual abilities of horses is necessary to ensure that this
species receives proper training, handling, management, and care. Author’s address: Equine Re-
search Foundation, PO Box 1900, Aptos, CA 95001; e-mail: EquiResF@aol.com; website: www.
equineresearch.org. © 2005 AAEP.

1. Introduction wild, they must cope with food and water of inconsis-
Traditionally, horses (Equus caballus) have rarely tent quality or unpredictable distribution, predators
been classified as intelligent, and even today, gaps in that change locations and habits, and a social system
knowledge, myths and misconceptions, and limited re- in which identities and roles of individuals must be
discovered and remembered.1 Domesticated horses
search affect how horses are understood or misunder-
may face even more potentially bewildering condi-
stood by the public, the horse industry, and even the
tions. In addition to dealing with similar situa-
scientific community. Common beliefs maintain that tions encountered in nature, many domesticated
horses have a brain the size of a walnut; horses do not horses must live in largely unsuitable environ-
think; horses are merely conditioned-response ani- ments, must suppress instincts while learning tasks
mals; horses cannot generalize; horses have no sense that are not natural behaviors, and must co-exist
of concept; horses are colorblind, have poor acuity and with humans who sometimes behave bizarrely, at
depth perception, and cannot transfer information least likely from an equine standpoint. Horses,
from one eye to another. both feral and domesticated, are faced with varied
In reality, horses manage not only ordinary daily conditions that require an assortment of learning
cognitive tasks but mental challenges as well. In the and perceptual capabilities.

NOTES

246 2005 ! Vol. 51 ! AAEP PROCEEDINGS

 IN-DEPTH: BEHAVIOR
Research is gradually providing insight into the
complex and fascinating nature of cognition and per-
ception in horses. This presentation will review
learning theory and key scientific studies to date on
these abilities. The findings are not only interest-
ing and useful for the education of clients but also
relevant to efficient veterinary interaction with
horses. Moreover, this information may be drawn
upon to enhance the training and behavior modifi-
cation of horses.

2. Habituation and Desensitization

Habituation is a learning process whereby, after
repeated exposure at the same intensity to an incon-
sequential stimulus, an animal becomes accustomed
and its reaction to that stimulus either diminishes
or disappears. A stimulus may be anything that Fig. 1. This horse is being positively exposed, without restraint, to
incites a response. Although this is an extremely a variety of visual, auditory, olfactory, and tactile stimuli. Stimuli
simple form of learning, it is, nonetheless, extremely include a bouncing beach ball, multicolored shiny mylar balloons,
important because it allows an animal to subcon- a bobbling umbrella, a plastic tarp, pylons, a barrel filled with
sciously filter out nonvital information in its envi- rocks dragged rapidly around the horse, and strangely behaving
humans. Photo credit: Evelyn B. Hanggi.
ronment and focus instead on what is significant.
From birth to old age, horses learn through habitu-
ation. In a new environment, many horses tend to
be much more reactive, at first paying attention to
any and all stimuli perceived by any of the senses. 3. Classical and Operant Conditioning and
However, when particular stimuli turn out to be Reinforcement
insignificant, the response diminishes. Horses, like most organisms, learn effortlessly
Desensitization is a process used to extinguish a through classical or Pavlovian conditioning, which
response to a stimulus in a sensitized or hypersen- occurs when initially unimportant stimuli or events
sitized animal. For instance, a horse may become are regularly paired with stimuli that initiate some
unusually fearful or sensitive to being bridled be- type of response. As a result, a new stimulus–re-
cause of rough handling of its ears or the bit banging sponse association forms in which an animal gives a
against its teeth. This leads to head shyness and response upon presentation of the original inconse-
the need for retraining. Such training will be more quential stimulus (now the conditioned stimulus),
problematic because the horse will attempt to avoid even though the unconditioned stimulus is no longer
the bridle; therefore, the handler will need to ap- available. Under this condition, the animal’s be-
proach the horse’s head in increments and retreat havior does not affect the occurrence of later events.
when necessary until the horse willingly accepts Horse trainers use classical conditioning regularly
normal, gentle bridling. This is an example of de- when they place a word onto a behavior, such as
sensitizing a horse that had become sensitized pairing the initially meaningless word “trot” with
through experience, but this process also works with the flick of a whip (previously associated with invok-
horses that are by nature overly sensitive. ing a flight response or pain) immediately before the
Good trainers take advantage of these learning horse changes gait during an upward transition.
abilities by exposing horses in a positive manner to Done consistently, it does not take long before the
horse responds with the appropriate action when
all sorts of sights, sounds, and contacts (Fig. 1).
given only the verbal cue.
When done correctly, horses become habituated and
Researchers, likewise, employ classical condition-
desensitized, even to potentially overwhelming ing to facilitate their experimental methodology.
stressors, and are much more capable of handling Hanggi2 uses the verbal conditioned stimulus “good”
novel events calmly. It is imperative, however, for to indicate correct responding during cognition test-
trainers to understand the principles behind the ing, thus informing the horse that a food reinforcer
practice before attempting to desensitize horses. is forthcoming.
Done incorrectly, for example, removing an alarm- To many a veterinarian’s dismay, horses quickly
ing stimulus too soon, can have the reverse effect— learn that the sight of a syringe is associated with
the horse becomes sensitized and more fearful. pain or discomfort. The unconditioned response
Sensitization, nonetheless, cannot be overlooked, be- (escape) then occurs whenever a horse that has not
cause it is the means of achieving responsiveness been trained to accept such handling catches sight of
and lightness to riding and handling aids. a syringe.3
AAEP PROCEEDINGS ! Vol. 51 ! 2005 247
 IN-DEPTH: BEHAVIOR
Stabled horses learn sounds associated with feed- on refinement: taking advantage of the principles
ing, such as the opening of hay room doors or grain of shaping and extinction by reinforcing the correct
poured into a bucket, or recognize visual cues, such and ignoring the incorrect, so that, over time, only
as the arrival of a feed person or vehicle. Within a the slightest pressure produces the desired action,
short time and to the frustration of owners, they which makes the human/horse partnership appear
display anticipatory behaviors, e.g., vocalizing, paw- effortless.
ing, or kicking stall doors, which, when reinforced, As popular as negative reinforcement is within
become conditioned behaviors. the industry, research employing avoidance condi-
Unlike classical conditioning, during operant con- tioning in horses is limited. In one study, ponies
ditioning, an animal manipulates its environment to that learned better in a positive reinforcement
obtain reinforcement— either positive (receiving single-choice point maze also learned better in a
something desired) or negative (removing something shock avoidance test. This indicates that some
unpleasant). Both types strengthen the relation- learning abilities are similar under positive and neg-
ship between a stimulus and the sought after re- ative reinforcement conditions.6 In another study,
sponse, so that when the same stimulus is given some horses performed better through positive rein-
again, there is an increased chance that the animal forcement, whereas others performed better when
will repeat the response.4 they were required to avoid an aversive stimulus.7
When a horse begins to learn the meaning of a Two other studies used sound or visual cues as
new stimulus, it will respond randomly, and through signals to avoid electric shocks for the purpose of
trial and error chance on the desired response. determining optimum number of conditioning ses-
Reinforcement of the response at the correct mo- sions8 and optimum number of trials per session.9
ment will cause the animal to repeat the behavior, In these tests, horses reached criterion in fewer tri-
albeit imperfectly at first. Horses excel at this type als when given only one avoidance session per week
of learning, especially when positive reinforcement instead of two to seven, but varying the number of
is available. For example, when presented with an trials per session did not affect their learning per-
environmental enrichment device called a foodball— formance. In other words, in these negative rein-
which dispenses pelleted feed as it rolls around on forcement studies, length of session was not critical,
the ground— horses will approach and investigate.5 but frequency was. This may not be true for all
Because of olfactory cues, they push the foodball types of training, as can be seen from anecdotal
with their noses or legs, causing it to roll and drop reports citing greater success with shorter sessions,
pellets. Most horses rapidly learn to manipulate especially with young horses.
the foodball in this manner, thereby receiving rein- Training for research purposes is predominantly
forcement and gaining control over this element of based on positive reinforcement. Horses learn to
their environment. It is believed that such control respond to a wide variety of stimuli in tests ranging
is important with regard to animal welfare, and from simple discrimination to concept learning, and
more attempts should be made to allow stabled training of basic experimental procedures flows eas-
horses some degree of instrumental control over ily. The use of positive reinforcement may be re-
their surroundings.1 stricted to rewarding correct test choices, or it may
A good understanding of positive reinforcement is be made the most of by chaining behaviors so that
very useful in working with equids. For instance, the horse, in effect, works completely on its own,
on hearing grain hitting a bucket, a horse may in- thereby minimizing inadvertent cueing by a han-
advertently kick the stall door, perhaps out of impa- dler. In many experiments, a handler leads a horse
tience. If a person then hurriedly feeds the horse into the testing area and turns it loose so that it may
(in a misguided hope of quieting it down), the kick- make a selection10,11; in others, a handler leads the
ing behavior will have been positively reinforced, horse all the way up to the test apparatus where it
and the person can quickly become a prompt for makes a choice.12–14 The latter runs the risk of the
kicking the stall door. Reinforced this way, it horse picking up unintentional signals from the han-
does not take long before the horse becomes an dler15 and is, therefore, open to criticism. To avoid
avid ruckus raiser capable of training humans such pitfalls, Hanggi2,16 –18 incorporates chaining
effectively. (linking a number of behaviors into a series) into
Operant conditioning is a horse training standard, experiments ranging from vision testing to categori-
and negative reinforcement has been the primary zation and concept learning (Fig. 2). Horses learn
means of shaping behaviors. Horses typically are to stand quietly in a “station” (a waiting area con-
trained to perform actions in order to avoid some- sisting of pylons with a bar spanning the front)
thing aversive. For example, under saddle, they during intertrial (Fig. 2A) and stimulus exposure
move forward when leg pressure from the rider is (Fig. 2B) intervals, to walk forward after the bar is
applied to both sides; on the ground, they yield their lowered and select a stimulus presented in one of
hindquarters when pressure is applied to a flank; two or more windows by touching it with its nose
they back up when pressure is applied to the bridge (Fig. 2C), to find a food reinforcer in a feeder located
of the nose; and they enter a trailer to avoid pressure on the bottom of the test apparatus, to walk away
from ropes or whips. The judicious trainer works from the apparatus (Fig. 2D) then halfway down the
248 2005 ! Vol. 51 ! AAEP PROCEEDINGS
 IN-DEPTH: BEHAVIOR

Fig. 2. Chaining of behaviors for re-

search purposes at the Equine Research
Foundation. Horses are trained to work
independently of a handler to minimize
inadvertent cueing. (B and C) The rel-
ative size concept study: for this horse,
“larger than” was the correct response.
Photo credit: Evelyn B. Hanggi.

length of the stable breezeway where they turn horses learned to load willingly, improper behaviors
around (Fig. 2E), and walk back into the station to disappeared, and these effects generalized to novel
await the next trial (Fig. 2F). Horses learn this situations. Although those of us who regularly use
chain within two or three training sessions and re- positive reinforcement and target training—for trailer
tain it indefinitely. loading (Fig. 3) and unloading, lifting feet for hoof care,
Practical applications of positive reinforcement have groundwork, standing quietly for grooming and veter-
been researched with respect to the use of these prin- inary handling, overcoming fear or resistance, and re-
ciples to facilitate trailer loading behavior in horses.19 search procedures—advocate this approach, more
Trailer loading the reluctant horse is a common prob- research and public demonstrations are needed to ed-
lem and can take hours to accomplish, with hazard to ucate horse handlers on the techniques and efficacy of
both horse and human. Resistant or frightened these methods. Ideally, trainers and handlers should
horses rear, pull back, kick, paw, and even fall over incorporate intelligent use of both positive and nega-
during the ordeal: these behaviors are reinforced tive reinforcement into a well-balanced program.
when owners fail to load them and give up. Tradi-
tional loading methods are based on negative rein- 4. Discrimination Learning
forcement (often with a measure of punishment Discrimination learning in horses has been reported
thrown in). Ferguson and Rosales-Ruiz19 found that, since the 1930s and is still regularly used today in
with positive reinforcement and target training, an array of tests. In discrimination tasks, horses
AAEP PROCEEDINGS ! Vol. 51 ! 2005 249
 IN-DEPTH: BEHAVIOR
(those relating to area) seem to be more discrim-
inable than other stimulus features to horses.26
Vigilant horse owners who have remarked on how
well horses find their way around in areas that they
have only visited infrequently support this claim.
Horses also react noticeably when objects in their
surroundings have been moved, indicating that they
recognize that something has changed spatially.
Nonetheless, horses are quite adept at discriminat-
ing visual stimuli: real life objects that have been
tested include buckets,11 doors,27 toys and photo-
graphs of toys,17 objects for everyday human use,18
abstract stimuli include striped patterns and col-
ors,28 and two-dimensional black figures.2,16,29
5. Discrimination and Visual Perception
Fig. 3. The author loading a horse into a trailer after positive
The ease with which horses discriminate stimuli facil-
reinforcement training. Traditional methods use whips, ropes,
negative reinforcement, and sometimes punishment, until the re-
itates not only cognition testing but also the measur-
luctant horse relinquishes. Horses do learn to load under negative ing of perceptual abilities, which in turn aids
conditions but some can become dangerous when stressed. Horses researchers in studying equine vision. For example,
that have learned to load through positive reinforcement do so in a study of depth perception, horses learned to dis-
eagerly, even with only a visual or verbal signal and even when criminate between two patterns using random-dot ste-
unrestrained. Photo credit: Jerry Ingersoll. reograms— one with no form visible and another
containing a square visible to individuals with stere-
opsis.30 Along with other depth perception studies,28
must learn that one stimulus, and not another, will this experiment provided evidence that horses pos-
result in reinforcement. That stimulus then begins sessed true stereopsis. Visual acuity was also tested
to control behavior, such that the horse acts in a in this manner. Horses were trained to choose be-
specific manner in the presence of one stimulus but tween stimuli made up of vertical black on white
not the other.20 Gardner21 found that horses could stripes of different widths.12 Discrimination testing
discriminate between a feed box covered with a continued until the horses could no longer differentiate
black cloth and boxes that were not. This simple the stimuli: results showed that a horse’s acuity is
discrimination was retained for over 1 yr. Horses "20/30 on the Snellen scale. This is not quite as good
showed a standard learning curve— errors de- as that of a “standard” human (20/20) but better than
creased as number of trials increased. However, dogs (20/50 or higher), cats (20/75 to 20/100), or even
these horses did not appear to generalize when the what the Department of Motor Vehicles requires for
cloth was located above or below the box. Another driving.
study showed that one horse was able to learn 20 Color vision in horses has also been tested
pairs of discriminations. This horse displayed the through the use of discrimination learning.14,31–33
ability of “learning to learn” by using a general so- This topic of great interest has yet to be satisfacto-
lution (one pattern in each pair was always re- rily resolved. Some studies showed that horses
warded) to more easily solve subsequent tests and could discriminate red and blue from gray, whereas
was able to retain 77.5% of the discriminations after others showed that they could discriminate not only
6 mo.22 The learning to learn phenomenon has these colors but green and yellow as well. How-
been noted in numerous other studies2,23–25 and is a ever, confounding factors, such as brightness, may
worthwhile tool in training. Too often, horses, es- have played a role to the degree that horses learned
pecially show horses, are limited to performing only to discriminate based not on color but on another
within a particular discipline. Thus, western plea- stimulus characteristic. Research recently com-
sure horses do not jump, and dressage horses rarely pleted by Hanggi and Waggonera confirmed that
set foot on a trail. This is unfortunate because such horses are color-deficient compared to humans with
restriction keeps the horse from learning about a trichromatic vision. Using discrimination tests,
great variety of stimuli, which creates an animal they showed that horses respond to color vision test-
that cannot deal with novel situations as comfort- ing in the same manner as some red/green color-
ably as one involved in a broad range of activities in deficient humans.
many different surroundings. Researchers, equine A myth that has surfaced repeatedly is that horses
welfare advocates, and good horse trainers agree cannot recognize with one eye what they have seen
that the more positive stimulation a horse experi- only with the other eye. This notion is used to
ences the more easily it learns in new situations and explain why horses startle at the same object when
the better adjusted it is in a variety of environments. viewed from different directions (such as when
Based on speed of acquisition and the extent to riding out and then coming back on a trail or revers-
which discriminations can be reversed, spatial cues ing directions in an arena). Anatomical examina-
250 2005 ! Vol. 51 ! AAEP PROCEEDINGS
 IN-DEPTH: BEHAVIOR
6. Generalization
Under stimulus generalization, a behavior previ-
ously conditioned to one stimulus transfers to other
similar stimuli. This adaptive trait permits an an-
imal to form associations with a wide range of stim-
ulus features rather than with only one element.
Under generalization testing, gradients (a series of
measured changes) are acquired that show how di-
rectly behavior is controlled by a given stimulus.
Responding is highest to the training stimulus
(which predicts reinforcement) yet still occurs, but to
a lesser degree, in the presence of stimuli possessing
certain features of the original stimulus.
Horses tested for stimulus generalization using cir-
cles showed symmetrical gradients: this is contrary
to gradients found in pigeons where generalization
Fig. 4. This Equine Research Foundation horse, while blind- tended toward the larger stimuli.34 Generalization
folded on the right eye, learned that the plus stimulus was cor- in horses was also examined using tactile stimuli—
rect. Evidence of interocular transfer (IOT) was shown when, repetitive tapping by solenoids along the horse’s
after the blindfold was switched to the other eye, it again cor- back—with results showing that behavior was effec-
rectly chose the plus. IOT was maintained regardless of stimu- tively controlled by the training stimulus.35 Horses
lus shapes, their location, or which eye was blindfolded responded most often to the training stimulus with
first. Photo credit: Evelyn B. Hanggi.
behavior decreasing as the stimuli were moved farther
away from the original location.
Lesson horses draw on generalization regularly,
understanding the assorted and inexact hand, leg,
tion has confirmed that the horse’s cerebral and seat cues from numerous riders of unequal skill
hemispheres do have a functional pathway for the and ability. On the other hand, generalization is
conveyance of information (belief contrary to this discouraged in dressage horses, which must discrim-
was another misconception), and a behavioral study inate highly precise cues from their riders.
showed that they do indeed have sufficient intero- Many horses could benefit from opportunities for
cular transfer.16 This study once again used mul- generalization. As mentioned earlier, horses in
tiple two-choice discrimination tests in which horses specific riding disciplines are frequently not allowed
were trained to respond to one stimulus and not to participate in activities other than what interests
another while blindfolded over one eye (Fig. 4). their riders. As a result, they go through mechan-
Once the discrimination was learned, the blindfold ical motions that rarely enhance any cognitive skill.
was switched to the other eye. Horses immediately Evidence of this can be seen in a recent study that
responded to the same stimulus, clearly indicating showed that, compared with horses involved in other
interocular transfer. disciplines, high-level dressage horses displayed the
Discrimination learning was also used to further lowest level of learning performance in simple
study why horses startle at objects they had appar- tests.36 It was hypothesized that because these
ently already seen. One intriguing hypothesis holds horses are trained to perform highly sophisticated,
that it is not a matter of recognizing the same object precise behaviors, riders give them minimal free-
during a return trip but more a matter of the object dom; therefore, they are inhibited from learning to
appearing different from another perspective. As learn or generalizing.
mentioned earlier, horses perform very well on spatial Horse trainers, from backyard to professional, can
discrimination tasks, and anecdotal evidence is vast enhance their horses’ generalization abilities by in-
with reference to horses noticing when objects in their corporating variety into their programs, both from
surroundings have been relocated. Therefore, it is the ground and in the saddle. This type of training
possible that horses do not always realize that an helps keep interest up, aids the horse when it comes
object is the same one when viewed from alternate across something new, and also gives human and
angles. Using children’s toys and a two-choice dis- horse an alternative to riding, especially during in-
crimination paradigm, Hanggib initially trained clement weather when horses are otherwise left to
horses to choose one of two objects, with the front of their own devices in stalls (Fig. 5).
both always positioned to the left. Once the discrim-
ination was learned, the objects were rotated front to 7. Observational Learning
right, front forward, front backward, upside down, etc. Horses are social animals most comfortable in the
The horses accurately chose objects presented in cer- company of other horses (Fig. 6). For many ani-
tain rotations but failed with others, indicating that mals, social interactions can facilitate the learning
recognition of rotated stimuli is good under some but of new behaviors. Thus, it would not be unusual
not all conditions. for horses to learn by observing others, and in
AAEP PROCEEDINGS ! Vol. 51 ! 2005 251
 IN-DEPTH: BEHAVIOR
onstrator conspecifics solve discrimination tasks
between buckets showed no sign of superior learning
over controls.25,38 Discrimination learning tests un-
der more rigorous conditions showed similar negative
findings, but observer horses did approach goal areas
more rapidly—leading to the assumption that some
learning had occurred.39 Observational learning also
did not seem to assist acquisition of a foot-press re-
sponse.40 Nevertheless, it is difficult to accept that
horses cannot learn by observation in any situation.
More likely, the proper experimental procedure has
yet to be developed. Perhaps a design more suited to
the nature of the horse might better reveal the extent
of observational learning in horses.
8. Categorization and Concept Learning
Many horse people believe, some even vehemently ar-
gue, that the learning abilities of horses do not go
Fig. 5. Nonriding activities can include generalization training, beyond the scope of associative learning and memory.
which teaches the horse to accept unusual and sometimes alarm-
Although a large amount of cognitive behavior can be
ing, stimuli in a relaxed manner. Photo credit: Jerry Ingersoll.
explained by these mechanisms, it is critical for the
well being of horses to study whether they possess
more advanced learning abilities. If the cognitive
fact, many horse owners believe this to be true. abilities of horses are misunderstood, underrated, or
McGreevy et al.37 reported that 72% of #1000 own- overrated, their treatment may also be inappropriate.
ers thought that abnormal behaviors were learned Equine welfare is dependent on not only physical com-
by observation. Abnormal or stereotypic behaviors, fort but mental comfort as well. Confining a thinking
including cribbing, weaving, head bobbing, pacing, animal in a dark, dusty stable with little or no social
and self-mutilation, are most often exhibited by sta- interaction and no mental stimulation is as harmful as
bled horses. Unfortunately, owners mistakenly as- providing inadequate nutrition or using abusive train-
sume that when groups of horses display the same ing methods. Therefore, it is in the interest of both
stereotypies it is because they have learned by ob- horses and humans to understand more fully the scope
serving one another. This leads to the detrimental of equine thinking.
act of removing a horse from any social contact. In comparison with the cognition work with other
In reality, the appearance of stereotypies in horses animals, little research into advanced equine learning
living near each other is more likely caused by ge- has been completed, which is astounding considering
netic relatedness or to the stress of existing in the the importance of horses to humans. Fortunately,
same, inappropriate environment. this is now changing as researchers design experi-
To date, there is no research supporting observa- ments that center on more complex cognitive skills.
tional learning in horses. Horses that observed dem- For example, the ability to categorize provides the
basis for substantial higher cognitive function.41
Categorization through the identification of similar
physical characteristics may involve stimulus general-
ization. Nicol1 noted that this should be functionally
valuable because it would allow animals to acquire
broad categories (food, predator, surroundings) and
react quickly in novel or unpredictable situations.
For example, endurance racehorses are confronted
with a great diversity of stimuli during their training
and competition. Rather than having to learn about
each object or event separately, they may make in-
stant classifications of new stimuli and adjust their
movements accordingly. Developing techniques to
incorporate categorization learning into everyday
training would undoubtedly be beneficial and would
give trainers another practical tool to use.
Categorization learning in horses has been exam-
Fig. 6. Being social animals, horses depend on conspecifics for ined in only a handful of studies. Using two-dimen-
safety and companionship. This is true for domesticated horses sional triangles, Sappington and Goldman29 found
as well as these free-roaming Mustangs in Nevada. Photo cred- that horses could discriminate triangles from other
it: Evelyn B. Hanggi. shapes; however, they were not able to provide con-
252 2005 ! Vol. 51 ! AAEP PROCEEDINGS
 IN-DEPTH: BEHAVIOR
complex level of conditional relations or relations
between relations or (2) a more simple level of un-
derstanding relative class concepts43 (e.g., bigger,
darker). This second level, although less compli-
cated, is still an effective measure of cognitive abil-
ity, because as Pepperberg and Brezinsky,43 when
paraphrasing Thomas,44 state:

Responding on a relative basis requires a subject

to compare stimulus choices and then derive and
use an underlying, more abstract (and thus gen-
eral) concept: in contrast, learning an absolute
stimulus value requires only that a subject form a
single association (p. 286).

In a discussion of concept learning in horses,

Fig. 7. This horse is making the correct choice in an Equine Nicol1 points out that even though Sappington and
Research Foundation categorization study. Stimuli were either Goldman29 refer to concept formation in their
open-centered or solid two-dimensional black figures of various horses, what they actually tested was categorization
geometric shapes, letters, or icons of U.S. states, people, or learning. Likewise, Flannery’s horses,13 in an
plants. Photo credit: Evelyn B. Hanggi. identity match-to-sample (sample and correct re-
sponse stimulus look identical) attempt, were not
tested using novel, unrewarded stimuli; therefore,
her findings can be explained by conventional asso-
clusive evidence of categorization with novel trian- ciative learning.
gles. Therefore, it is possible that the horses Hanggi18 studied the ability of horses to form con-
responded correctly because of associative learning cepts based on relative size (larger than, smaller
processes.1 The ability to categorize was more ev- than). After learning to respond to the larger of
ident in a study by Hanggi,2 in which horses were two stimuli for six sets of two-dimensional training
trained to discriminate between two-dimensional exemplars, two horses were tested for size transpo-
black figures with open centers and solid black fig- sition that used novel larger and smaller stimuli as
ures (Fig. 7). Horses learned the initial discrimi- well as three-dimensional objects (sizes: large, me-
nations by operant conditioning but then acquired dium, small, tiny; Fig. 2). Horses chose the larger
subsequent discriminations with fewer errors, indi- of two (or more) stimuli, regardless of novelty or
cating that they had learned to learn. Evidence of dimension. A second phase of this study tested
the formation of a category came with the introduc- “smaller than” and produced the same results, that
tion of novel stimuli. Horses immediately chose the is, the horse selected the smaller stimuli based on
open-center stimuli in most instances, which would relative size. Moreover, the horses generalized
have been at chance levels had they been function- across situations that varied from black two-dimen-
ing only by associative learning. sional shapes to three-dimensional objects of differ-
Interestingly, under certain circumstances, ent materials and colors, including yellow foam
horses have difficulty transferring learned stimuli to balls, green plastic flowerpots, and red PVC connec-
new tasks.42 In one case, horses failed to use fa- tors. All horses responded as if they had developed
miliar stimuli (a striped board or white bucket pre- a concept of relative size, marking this study as the
viously used as cues for lever pressing) to solve a first to provide evidence of relative class concept
novel, unrelated task (maze). This may have been formation in equids.
because of experimental design and insufficient gen- This is not to say that horses possess the same
eralization training before the start of the experi- conceptualization abilities as humans, nonhuman
ment or this specific type of transfer may indeed be primates, or other so-called advanced species, but it
difficult for horses in general. Further study is is an indication that they possess more cognitive
needed before conclusions can be reached regarding ability than what was known. Much more focus
this detail of equine learning. should be placed on these types of abilities, not just
Another topic barely discussed with respect to to satisfy scientific curiosity but also for practical
horses, except from a negative perspective, is concept purposes. As with the application of categorization
formation. Unlike categorization, where stimuli re- learning and generalization ability to training pro-
semble one another physically, conceptualization in- cedures, simple concept learning may be applied to
volves responding to certain stimuli because they training and handling. Equine welfare with re-
represent the same idea, regardless of whether they spect to care and management may also benefit
physically resemble one another. when more is discovered about abstract concept for-
Concept learning may involve either absolute or mation. It has been suggested that, if horses do not
relational concepts, with the latter divided into (1) a have a concept of “sameness,” we cannot expect
AAEP PROCEEDINGS ! Vol. 51 ! 2005 253
 IN-DEPTH: BEHAVIOR
them to instantly accept new situations that are, to have the capability of seeing colored objects in their
us, just the same as previous ones. For example, environment, even when those objects are nearly
humans classify all horse trailers as the same be- camouflaged and even under conditions when their
cause of their function and so might expect horses to reported visual acuity would not suffice. Rather
readily enter any type after they have been trained than considering the whole horse, the vision re-
with just one.1 According to Nicol,1 horses may not searchers made judgments based on only a few fea-
possess such a shared functional concept and will tures and consequently came to the wrong
have to learn to enter each type individually. This conclusion. When all details of the accident were
idea has merit in that it should make handlers con- examined, it seemed much more likely that the
sider the environment more from the horse’s point of horse crashed into the fence because of a combina-
view. However, at least in the trailer example, tion of equine behavior and rider error.
there is more to the issue than concept capability.
In many instances, horses refuse to enter trailers 10. Conclusion
because of poor training, negative experiences, or In the not too distant past, little consideration was
lack of opportunity to generalize. Horses that have given as to why horses behaved as they did or what
been properly exposed to myriad stimuli ranging possibilities existed to ensure adequate care and
from stables to loafing sheds, one-horse trailers to welfare. However, during the past decade, an ex-
stock trailers, crowded places to wide-open spaces, plosion of sorts has occurred within the horse indus-
and chaotic traffic to quiet countryside are remark- try. Scientific conferences, research articles, horse
ably adaptable in novel situations. On the other expositions, clinics, the Internet, television, maga-
hand, horses cloistered in isolated stalls and only zines, books, videos, and worldwide tours have
brought out for repetitive training are apt to react served the public by making available educational
adversely when novelty arises, and many tend to- prospects never before seen on such a large scale.
ward nervousness, flightiness, and neurotic behav- From skilled horse handlers to wide-eyed novices,
ior. Perhaps it is not a matter of possessing a everyone has the opportunity to advance their
cognitive ability or not, rather it is a function of knowledge about equine cognition, behavior, train-
upbringing, training, and opportunity to learn. ing, and care. Unfortunately, as is human nature,
9. More Than the Sum of the Parts some equine authorities take advantage of those
eager for information, creating a persona of near
As is generally the case in research, only particular thaumaturgy. It is up to the individual to differen-
aspects of an organism are studied. This occur- tiate between the sincere and the artificial, to search
rence arises from the interests of the researcher or for truth among unsubstantiated declaration, and to
some other entity and is not, in itself, a problem become an eclectic in the world of horses.
unless the item of interest overshadows everything Research into equine cognition and perception has
else. Take, for example, the attention given to made noteworthy advances, and greater interest is
equine color vision and visual acuity. Substantial now being paid to the improvement of training
effort has been made to discern what and how the methodology and management. Nonetheless,
equine eye sees. This is admirable, that is, until much more still needs to be learned before scientists
assertions are made regarding equine behavior that and laypersons alike can make unequivocal claims
may or may not have anything to do with vision. as to what it is to be a horse. Studies combining
For example, in one case, a small group of vision equine learning, perception, and behavior are the
researchers claimed that a racehorse crashed into a next step in understanding this remarkable animal.
green fence because of the color, which according to
their research could not be seen by horses. How- The Equine Research Foundation is supported by
ever, when horses were observed behaving natu- grants and public donations. I thank Betty and
rally, as well as under test conditions, they easily Meinrad Hanggi and Jerry Ingersoll for their enthu-
avoided green objects even while moving at high siastic support and efforts. Research conducted at
speed.c This led Hanggic to further test this phe- the ERF is assisted by student volunteers and
nomenon by placing green objects of various shapes participants of the ERF learning vacations and
and sizes against a similarly colored green back- internships.
ground (a plastic tarp) set against green trees and
observing whether the horses could locate them.
References and Footnotes
A horse was stationed 18 feet from the background
and then allowed to walk forward and make contact 1. Nicol CJ. Learning abilities in the horse. In: Mills DS,
McDonnell SM, eds. The domestic horse: the evolution, de-
with the object with a nose-touch. The horse had velopment, and management of its behavior. Cambridge, MA:
no difficulty noticing the stimuli, regardless of their Cambridge University Press, 2005;169 –183.
location on the background. Most notably, it could 2. Hanggi EB. Categorization learning in horses (Equus cabal-
still find an object that was only 1 $ 2 $ 5/8 in and lus). J Comp Psychol 1999;1113:243–252.
3. Houpt KA. Learning in horses. In: The thinking horse.
covered with the same material as the background. Guelph, Canada: Equine Research Centre, 1995;12–17.
This simple test showed that, even though horses 4. Tarpy RM. Basic principles of learning. Glenview, IL:
may not pass certain color vision tests, they still Scott, Foresman and Co., 1975.

254 2005 ! Vol. 51 ! AAEP PROCEEDINGS

 IN-DEPTH: BEHAVIOR
5. Winskill LC, Waran NK, Young RJ. The effect of a foraging 25. Baer KL, Potter GD, Friend TH, et al. Observation effects
device (a modified Edinburgh Foodball) on the behaviour of on learning in horses. Appl Anim Ethol 1983/84;11:123–
the stabled horse. Appl Anim Behav Sci 1996;48:25–35. 129.
6. Haag EL, Rudman R, Houpt KA. Avoidance maze learning 26. Nicol CJ. Equine learning: progress and suggestions for
and social dominance in ponies. J Anim Sci 1982;50:329 – future research. Appl Anim Behav Sci 2002;78:193–208.
335. 27. Mader DR, Price EO. Discrimination learning in horses:
7. Visser EK, van Reenen CG, Schilder MBH, et al. Learning effects of breed, age and social dominance. J Anim Sci 1980;
performances in young horses using two different learning 50:962–965.
tests. Appl Anim Behav Sci 2003;80:311–326. 28. Timney B, Macuda T. Vision and hearing in horses. J Am
8. Rubin L, Oppegard C, Hintz HF. The effect of varying the Vet Med Assoc 2001;218:1567–1574.
temporal distribution of conditioning trials on equine learn- 29. Sappington BF, Goldman L. Discrimination learning and
ing behavior. J Anim Sci 1980;50:1184 –1187. concept formation in the Arabian horse. J Anim Sci 1994;
9. McCall CA, Salters MA, Simpson SM. Relationship between 72:3080 –3087.
number of conditioning trials per training session and avoid- 30. Timney B, Keil K. Local and global stereopsis in the horse.
Vision Res 1999;39:1861–1867.
ance learning in horses. Appl Anim Behav Sci 1993;36:291–
31. Grizmek VB. Versuche über das Farbsehen von Pflanze-
299.
nessern. Tierpsychology 1952;9:23–39.
10. McCall CA. The effect of body condition of horses on dis-
32. Pick DF, Lovell G, Brown S, et al. Equine color perception
crimination learning abilities. Appl Anim Behav Sci 1989; revisited. Appl Anim Behav Sci 1994;42:61– 65.
22:327–334. 33. Smith S, Goldman L. Color discrimination in horses.
11. Sappington BF, McCall CA, Coleman DA, et al. A prelimi- Appl Anim Behav Sci 1999;62:13–25.
nary study of the relationship between discrimination rever- 34. Dougherty DM, Lewis P. Stimulus generalization, discrim-
sal learning and performance tasks in yearling and 2-year-old ination learning, and peak shift in horses. J Exp Anal Be-
horses. Appl Anim Behav Sci 1997;53:157–166. hav 1991;56:97–104.
12. Timney B, Keil K. Visual acuity in the horse. Vision Res 35. Dougherty DM, Lewis P. Generalization of a tactile stimu-
1992;32:2289 –2293. lus in horses. J Exp Anal Behav 1993;59:521–528.
13. Flannery B. Relational discrimination learning in horses. 36. Hausberger M, Bruderer C, Le Scolan N, et al. Interplay
Appl Anim Behav Sci 1997;54:267–280. between environmental and genetic factors in temperament/
14. Macuda T, Timney B. Luminance and chromatic discrimi- personality traits in horses (Equus caballus). J Comp Psy-
nation in the horse (Equus caballus). Behav Processes 1999; chol 2004;118:434 – 446.
44:301–307. 37. McGreevy PD, French NP, Nicol CJ. The prevalence of ab-
15. Pfungst O. Clever Hans (the horse of Mr. Von Osten). normal behaviours in dressage, eventing and endurance
Thoemmes Continuum, Bristol, 2000. horses in relation to stabling. Vet Rec 1995;137:36 –37.
16. Hanggi EB. Interocular transfer of learning in horses 38. Baker AEM, Crawford BH. Observational learning in
(Equus caballus). J Equine Vet Sci 1999;19:518 –523. horses. Appl Anim Behav Sci 1986;15:7–13.
17. Hanggi EB. Can horses recognize pictures?, in Proceedings. 39. Clarke JV, Nicol CJ, Jones R, et al. Effects of observational
Third International Conference on Cognitive Science 2001; learning on food selection in horses. Appl Anim Behav Sci
52–56. 1996;50:177–184.
18. Hanggi EB. Discrimination learning based on relative size 40. Lindberg AC, Kelland A, Nicol CJ. Effects of observational
concepts in horses (Equus caballus). Appl Anim Behav Sci learning on acquisition of an operant response in horses.
2003;83:201–213. Appl Anim Behav Sci 1999;61:187–199.
19. Ferguson DL, Rosales-Ruiz J. Loading the problem loader: 41. Nicol CJ. Farm animal cognition. J Anim Sci 1996;62:
the effects of target training and shaping on trailer-loading 375–391.
behavior of horses. J Appl Behav Anal 2001;34:409 – 424. 42. McCall CA, Salters MA, Johnson KB, et al. Equine utiliza-
tion of a previously learned visual stimulus to solve a novel
20. McCall CA. A review of learning behavior in horses and its
task. Appl Anim Behav Sci 2003;82:163–172.
application in horse training. J Anim Sci 1990;68:75– 81.
43. Pepperberg IM, Brezinsky MV. Acquisition of a relative
21. Gardner LP. The responses of horses in a discrimination
class concept by an African Gray parrot (Psittacus erithacus):
problem. J Comp Psychol 1937;23:13–33. discrimination based on relative size. J Comp Psychol 1991;
22. Dixon J. The horse: a dumb animal?. . . Neigh! Thorough- 105:286 –294.
bred Rec 1970;1655–1657. 44. Thomas RK. Evolution of intelligence: an approach to its
23. Fiske JC, Potter GD. Discrimination reversal learning in assessment. Brain Behav Evol 1980;17:454 – 472.
yearling horses. J Anim Sci 1979;49:583–588.
24. McCall CA, Potter GD, Friend TH, et al. Learning abilities a
Hanggi EB, Waggoner TL. Unpublished data. 2005.
b
in yearling horses using the Hebb-Williams closed field maze. Hanggi EB. Unpublished data. 2005.
c
J Anim Sci 1981;53:928 –933. Hanggi EB. Unpublished data. 2005.

AAEP PROCEEDINGS ! Vol. 51 ! 2005 255