"Until recently,
horses have been estimated to have average intelligence at
best," said Evelyn B. Hanggi, MS, PhD, president and, along with
Jerry Ingersoll, co-founder of the Equine Research Foundation (ERF)
in Aptos, Calif., during her review of literature on equine
cognition and perception at the 2005 American Association of
Equine Practitioners Convention. "Even today, gaps in knowledge,
myths and misconceptions, and limited research affect how horses
are understood or misunderstood by the public, the horse
industry, and even the scientific community. Luckily, both
professionals and lay people are becoming more interested in
equine learning/cognition."
Why is it important to know how and how
well horses can think and learn? "If the cognitive abilities of
horses are misunderstood, underrated, or overrated, their
treatment may also be inappropriate," said Hanggi. "Equine
welfare is dependent on not only physical comfort, but mental
comfort as well."
In other words, if we understand how a
horse thinks and learns, we can avoid mentally abusing a horse
by not expecting him to learn what he is unable to learn.
Conversely, we can also provide him with proper stimulation to
learn that which he is able to learn quickly and easily.
Hanggi described various types of learning
and perception in horses, and she reviewed research describing
each one. These include:
- Habituation,
- Desensitization,
- Classical or Pavlovian conditioning,
- Operant conditioning,
- Discrimination,
- Visual perception,
- Generalization,
- Observational learning,
- Categorization, and
- Concept learning.
She also provided several videos of horses
at the Equine Research Foundation in research environments and
training situations to show how positive reinforcement can help
horses quickly learn a task.
Habituation/Desensitization
Habituation is a very simple form of
learning, said Hanggi. It consists of a horse initially
responding to a stimulus, then learning to filter it out and
responding less or not at all. This is helpful because it allows
the horse to focus on just what is important in its environment,
not every little thing. Desensitization is similar, but it is a
process that consists of extinguishing a response in a horse
that is or has become overly sensitive to a stimulus through
experience.
"Good trainers take advantage of these
learning abilities by exposing horses in a positive manner to
all sorts of sights, sounds, and contacts," she said. "When done
correctly, horses become habituated and desensitized, even to
potentially overwhelming stressors, and are much more capable of
handling novel events calmly."
Classical and Operant Conditioning
Classical or Pavlovian conditioning occurs
when a horse learns to associate some unimportant stimulus with
one that generates a response, such as learning that the rattle
of the feed cart means dinner is coming, she said. Eventually
the horse responds to the originally unimportant stimulus as if
it were the important one. However, he has no control over these
events; this is different from operant conditioning, in which
the horse has or perceives he has control over the outcome. One
example of operant conditioning might be a horse learning to
roll around a food ball so the food (his reward) falls out of
the holes for him to eat.
"Operant conditioning is a horse training
standard, and negative reinforcement has been the primary means
of shaping behaviors," Hanggi said. "Horses typically are
trained to perform actions in order to avoid something aversive
(such as moving away from pressure)."
Positive and Negative Reinforcement
At this point, she described the principles
of positive reinforcement (rewarding of desired behavior) and
negative reinforcement (removal of something somewhat unpleasant
such as leg pressure) as they relate to horse learning. She
noted that while various studies have found conflicting results
regarding which is more effective for horses, training for
research purposes is primarily based on positive reinforcement.
Little work has been done on negative reinforcement in horses.
"Ideally, trainers and handlers should incorporate intelligent
use of both positive and negative reinforcement into a
well-balanced program," she recommended.
"A good understanding of positive
reinforcement is very useful in working with equids," she went
on. "For instance, on hearing grain hitting a bucket, a horse
may inadvertently kick the stall door, perhaps out of
impatience. If a person then hurriedly feeds the horse (in a
misguided hope of quieting it down), the kicking behavior will
have been positively reinforced, and the person can quickly
become a prompt for kicking the stall door. Reinforced this way,
it does not take long before the horse becomes an avid ruckus
raiser capable of training humans effectively.”
"People often confuse food reinforcement
with treat giving, which can lead to invasion of personal space,
biting, etc.," she commented. "Teaching a horse to look away or
straight ahead for food reinforcement is the key toward
establishing your personal space.
"Positive reinforcement teaches the horse to become an active
participant seeking the right answer," Hanggi explained. She
offered several videos of positively reinforced target training
to overcome problems like head shyness when bridling and when
having a facial cut medicated.
Discrimination
Discrimination learning involves positive
reinforcement at the Equine Research Foundation; the horse
learns that one stimulus, and not another, will result in
reinforcement, said Hanggi. One example might be the horse
learning to select a particular bucket color, size, shape, etc.,
rather than another consistently when presented with both. This
type of learning in horses has been reported since the 1930s,
she noted.
An interesting finding from discrimination
tests has been that many horses will "learn to learn," applying
successful behaviors from one trial to another with a resulting
increase in success. Hanggi noted that this is a worthwhile tool
in training, and it supports the use of cross-training to
improve a horse's handling of novel situations (more on this
shortly).
Owners might wonder what types of things a
horse is able to visually discriminate, such as colors or
shapes. Hanggi said studies have shown that horses do possess
some degree of stereopsis (vision melding the images from both
eyes to gain depth perception), have a visual acuity of about
20/30, and probably have red/green colorblindness.
It has often been said that horses can't
recognize with one eye things they have only seen with the other
eye, said Hanggi. "This notion is used to explain why horses
startle at the same object when viewed from different directions
(such as when riding out and then coming back on a trail or
reversing directions in an arena)," she explained. However, an
ERF study has shown this to be untrue. Horses can easily
transfer information between their eyes; this is called
interocular transfer.
According to some research, spatial cues
(when objects move around) seem to be easier for horses to
discriminate than some other cues. Also, she reported that
horses recognize rotated objects in most orientations except
upside-down.
Another ERF study "showed evidence that
horses can indeed recognize photos of objects, and objects from
photos," Hanggi reported. "We are also looking at a long-term
study on facial expressions."
Generalization
This type of learning means the horse
transfers the response he has to one stimulus to other similar
stimuli. For example, lesson horses learn to respond to leg
pressure even when it is applied in various locations and to
different degrees by different riders, Hanggi explained.
However, generalization is discouraged in some highly trained
horses, such as dressage horses.
"Many horses could benefit from
opportunities for generalization," she opined. "Horses in
specific riding disciplines are frequently not allowed to
participate in activities other than what interests their
riders. As a result, they go through mechanical motions that
rarely enhance any cognitive skill. Evidence of this can be seen
in a recent study that showed that, compared with horses
involved in other disciplines, high-level dressage horses
displayed the lowest level of learning performance in simple
tests. It was hypothesized that because these horses are trained
to perform highly sophisticated, precise behaviors, riders give
them minimal freedom; therefore, they are inhibited from
learning to learn or generalizing."
She suggested that horses and riders would
benefit from adding variety to their riding and ground work
programs.
Observational Learning
While nearly three-quarters of owners in
one study believe that horses can learn negative behaviors from
each other (such as cribbing), Hanggi said as yet there is no
research supporting this theory. "In reality, the appearance of
stereotypies in horses living near each other is more likely
caused by genetic relatedness or to the stress of existing in
the same, inappropriate environment," she said.
"Nevertheless, it is difficult to accept
that horses cannot learn by observation in any situation," she
commented. "More likely, the proper experimental procedure has
yet to be developed."
Categorization and Concept Learning
"Categorization and concept learning are
the highest levels of learning that have been tested in horses,"
said Hanggi. One example of categorization might be horses
learning to pick open-center shapes over solid ones regardless
of the actual shape or its size, as occurred in one ERF study.
"The ERF has also shown, for the first time, that horses are
capable of some degree of conceptualization," said Hanggi. "In
one groundbreaking experiment, horses demonstrated that they
could solve problems using the concept of relative size. They
were trained to always choose the larger (or smaller for one
horse) object regardless of whether or not it was previously
correct. This means that when a medium-sized circle was paired
against a small circle, the medium one was correct. However,
when the medium circle was paired against a large circle, the
large (and not the previously correct medium) circle was
correct. This relative class concept transferred to other shapes
such as triangles, hexagons, and icons of U.S. states, as well
as to three-dimensional real-life objects such as foam balls and
plastic flower pots.
"This is not to say that horses possess the
same conceptualization abilities as humans, non-human primates,
or other so-called advanced species, but it is an indication
that they possess more cognitive ability than what was known,"
she stated. "Much more focus should be placed on these types of
abilities, not just to satisfy scientific curiosity, but also
for practical purposes. Studies combining equine learning,
perception, and behavior are the next step in understanding this
remarkable animal."
Research and Training:
If You Can Think It, You Can Train It
Participants in our learning and riding
vacations at the Equine Research Foundation (ERF) have a strong
desire to gain a better understanding of not only how the
horse’s mind works but also how to apply our scientific findings
to everyday training and handling. At first glance, it might not
seem that science and training are closely related but, in
reality, they go hand in hand.
From daily contact with our horses we know how easily they can
be taught through the more simple forms of learning, such as
habituation or desensitization and operant conditioning using
positive or negative reinforcement. But our research at the ERF
shows that horses are capable of much more. Most exciting are
our findings that show how well horses generalize and how they
are able to think using categories and concepts, at least to
some degree.
This is new to the horse world. In the
past, horses were trained mostly through negative reinforcement,
e.g., horses learned a behavior by moving away from a somewhat
unpleasant stimulus, such as leg pressure. More recently,
positive reinforcement training has started to draw attention.
Trainers are discovering that horses are extremely interested in
learning behaviors that result in reward and, when done
properly, training proceeds rapidly and nearly effortlessly. At
the ERF, we supplement these methods by developing our horses’
abilities to generalize and to use simple concepts.
With this, we are able to improve our
training methods so that they are geared more towards how the
horse perceives the world, rather than the typical
anthropocentric tactic. As we incorporate each new ERF finding,
we notice remarkable changes in our horses. Their attention and
focus improves dramatically. They learn to learn so that new
tasks are picked up more quickly and easily. They deal with
trailer loading, vet and farrier care, under saddle work,
groundwork and just about any other handling without stress.
They handle new experiences better and startle much less. Their
desire to be around us becomes noticeably stronger and the
human/horse bond grows into something amazing. Our horses seek
out the correct answers eagerly and actually appear to enjoy
what they are doing with us. Likewise, we have fun working with
them. Indeed, when training is enjoyed, it is better for both
horse and human and creates a relaxed state of mind.
Anyone who trains horses takes advantage of
science and psychology, whether intentionally or not. Those who
excel at training also possess a solid understanding of the
principles behind the practice and that makes the difference
between a mediocre training experience and one that ends in
great success.
