maintained by
The Horse Magazine
PO BOX 349,
PAKENHAM, VICTORIA,
AUSTRALIA 3810.
PH: (03) 5942 7447
FAX: (03) 5942 7556
Email Us
ALL MATERIAL
APPEARING ON THIS
SITE IS COPYRIGHT ©
Reproduction in whole
or in part without
permission is not
permitted.
PAVLOV’S PRINCIPLE
by Andrew McLean
Relaxation and attentiveness can only occur if the horse responds to
light signals. During training the horse learns to respond from light
aids if they are offered just before and overlapping stronger motivating
pressure. When training a new signal, the new signal should happen
just before and during the light aid.
Pavlov’s principle is all about training the horse to operate
from light aids, including seat, weight and positional cues. There
is a definite science about training horses and other animals to operate
from previously neutral signals, and Ivan Pavlov (1849 – 1936)
was the first to describe it.
He was a Russian scientist and in fact wasn’t searching for any behavioural principles when he discovered and revealed the process of acquiring signals. This process is now known as Classical or Pavlovian conditioning. Pavlov was doing some rather gory experiments on a dog with a glass plate sewn over a large hole cut right through its ribcage to its stomach so he could see the digestive process first hand. He gave the dog a meat extract so he could see what goes on in the animal’s stomach during salivation and digestion. During the course of the experiments however, he soon noticed that the dog began to salivate earlier on in each session – the dog was beginning to ‘anticipate’ the food.
Now most of us would be happy to explain the dog’s anticipation as simply that, but Pavlov was determined to delve deeper. He wanted to know what ‘anticipation’ actually was - to uncover the mechanistic principle behind anticipation. In humans and perhaps in some animals as well it can involve some kind of mental visualisation of the anticipated event, but is this always the case? And does it need to be so? Pavlov found that there were some strict rules that govern the way an animal learns to associate a previously neutral stimulus with a particular response. He found that the timing of when the light signal appears in relation to the already known stimulus is critical.
TIMING OF ASSOCIATIONS
Pavlov showed that new cues must be given just before or during an
inborn or previously learned signal or aid. For example, if the rider
wishes to train the voice aid “whoa”, the word should occur
just before and overlapping the application of the rein signal for ‘stop’.
The further the word occurs after the rein aid, the less the horse
learns the voice aid. The rider must also remember that as the horse
does not actually understand what “whoa” means, but rather
learns it as a cue, it must always be delivered with the same tone
and pitch. Shouting “whoa” won’t make the horse stop
more quickly, it is only the learned version of the voice aid that
is meaningful to the horse.
In dressage voice aids are not used in
competition. Instead, as the horse learns to respond from lighter rein
and/or leg aids, the horse becomes increasingly aware of the associations
of seat and weight that naturally occur with rein and/or leg aids if
the rider’s position is correct. These responses are learned
through classical conditioning as the seat and weight aids occur just
before and during the rein and /or leg aids. Similarly during training
the horse to lunge, the horse learns to respond to voice aids. However
the fact that the trainer needs to keep the lunge whip in his hand
is testimony to the fact that classical conditioning is but a thin
veneer in training. Voice, seat, weight and position associations are
easily forgotten and need reminding with the primary aids of rein and/or
leg and in the case of lunging, the whip.
Pavlov’s work suggested that Classical conditioning doesn’t
need any kind of visualisation or comprehension in the animal of an
anticipated event in order to work - it simply needs the correct order
of presentation. That’s why Classical conditioning can be seen
in worms and houseflies as well as horses and humans. If the sound
of running water makes you want to go to the toilet, that’s classical
conditioning. As you respond to traffic lights and other road signals
yet you aren’t consciously aware of what you are doing, you just
do it – that’s Classical conditioning. The life of all
animals is full of associations that have formed and will form between
various signals and events. Acquiring signals that predict events makes
life predictable and controllable for animals. It evolved to improve
the efficiency of the animal’s interaction with its environment.
Any environmental event that coincides with a known signal will quickly
become incorporated as a new cue. So, the sound of rustling bushes
which precedes the appearance of a predator - and triggers the flight
response in the horse - quickly becomes a cue for running away.
PREDICTABILITY
Horse training relates to this notion of predictability in a very definite
way. The reason we want to place all of the horse’s trained responses
under the control of light aids isn’t just for our convenience
and laziness – it is for the horse’s mental well being.
The horse needs unobtrusive, pain free signals for all its movements
in hand and under saddle. Good coaches and trainers have long known
the importance of lightness of the aids as well as the fact that if
the horse needs stronger pressure to motivate him to do something,
then these pressures must always be preceded by light aids. But it
was the writings of a man called Piet Wiepkema, a Dutch cognitive scientist,
which first brought me to understand the relationship between an animal’s
well-being and the nature and consistency of the signals that it encounters
in its life.
Take a moment to contemplate this – any animal’s existence involves giving and receiving signals and/or ‘pressures’ from its environment. The signals/’pressures’ it issues to its environment in order to procure benefits for itself are either inborn ones or learned ones, and the ones it receives from its environment are either from the physical or the behavioural world (including signals from its rider or handler). Animals have evolved the ability to offer and learn to respond to mild unobtrusive signals so that they don’t have to endure a life of painful or unpredictable events such as the sudden attack by another horse during a squabble over resources such as food or mates as well as to predict predatory attacks. For example, horses soon learn that before another horse attacks, it lays its ears back.
CONSISTENCY
Animals are thus able to learn signals that surround all events that
not only predict nasty things but also nice ones too, such as the arrival
of the person carrying a bucket – it means food. Wiepkema showed
that how often a particular signal consistently predicts a particular
response is directly proportional to the amount stress in an animal.
If the signal always leads to the same response, the animal is relaxed
in its response to the signal. Think of your own life. What makes you
calm (or not) as an adult is that you have (or have not) found ways,
generally using language, to control your response to others and to
control the behaviour of others. All organisms need to be able to make
their behavioural world predictable. The less predictable and controllable
their world is, the more stress and tension they show.
In horse training, predictability comes through the horse learning
to respond in the same way from a light aid for each response (stop,
go, turn and leg-yield). Responding in the same way means responding:
immediately to the light aid,
in a self-maintained rhythm and tempo,
with self-maintained straightness,
with self-maintained contact and outline,
with impulsion,
with all of the above properties everywhere and every time.
So when you look at the above qualities of the rein and leg aids, you
will appreciate that each response requires the development of many
properties. These properties need to be trained one by one as you will
see in a later article.
This is why correct horse training has always focussed on producing
a consistent set of responses each time an aid is given rather than
a random assortment of various incorrect responses. The German training
scale is the best known of mankind’s attempts to train consistency
of outcome in horse training. What is not well known in any equitation
discipline is that problem horses are a result of defects in consistency
of outcome from the aids. Instead, horsepeople describe a horse’s
training in terms of its ‘will to please’ rather than its
reaction to the aids. Horses are frequently described as ‘naughty’, ‘dirty’, ‘dumb’ or ‘hot’ rather
than using terms that describe what the horse’s legs do or do
not do in response to the aids.
SIGNAL PRIORITY
It is important in horse training to recognise that there is a priority
in training signals. At the very earliest stages in horse training,
the horse learns to respond from pressures, such as lead rein pressure
for leading, and under saddle pressure from both legs means go forward
while pressure from the reins means slow. However, good trainers ensure
that at the very beginning of each rein or leg pressure, that there
is a light version of that particular pressure. This has been known
for centuries and is described at length in classical training literature.
This light aid therefore is the very first cue that the horse learns
in-hand and under-saddle for go, stop, turn and leg-yield. The horse
learns through Pavlov’s principle, (Classical conditioning) to
respond from those light versions of the pressures.
Furthermore, during this training the horse also learns, again through
Pavlov’s principle to respond from associated seat, weight and
position cues because they occur just before and during the light rein
and leg aids. Horses learn these aids readily and unfortunately sometimes
well before the horse has thoroughly learned about the pressures that
actually enforce responses. Therefore, learning to respond from pressures
and light aids should always precede any reliance on seat and position
cues. Too many people rely on seat and position aids and forget to
either establish or maintain the underlying foundations of rein and
leg signals
. When the response to the seat and position aids begin to fail or even take too long to work and the horse no longer responds as he should, he soon forgets these aids entirely, as Pavlov predicted. Pavlov found that a conditioned response will be repressed if the stimulus proves “wrong” too often. If the seat does not produce a reliable response, the horse will stop responding to it. The same is true for the leg and rein aids – if they don’t work, the stronger motivating pressure should follow. It’s as if the horse is saying: “Please don’t use a force 6 pressure, I’ll do it from the light aid.’’ Using the right amount of pressure is a vital skill in horse training – not too much and not too little. Problems also occur if the rider maintains the mild pressure of the light aid when the horse has already responded. The horse desensitises to the light aid.
LOSSES OF PREDICTABILITY
However the problem with the horse that has become desensitised to
the aids for whatever reason is not just that he loses his response
to light signals. There is a big price that is often paid for this
and it is called conflict behaviour. Conflict behaviour incudes flight
response behaviours (i.e. fast ones) such as shying, bolting, bucking,
rearing and leaping. It also accompanies associated health and welfare
issues that include worsening colic attacks, immune suppression, hormonal
changes and poor and ‘stringy’ body condition. Conflict
behaviour arises from the stress that occurs due to losses, from the
horse’s viewpoint, of predictability and controllability of its
behavioural world. The horse is trying to run away from the stressful
situation.
Conflict behaviour may also arise when the trainer does not regularly
target a consistent set of responses from an aid. When all of the properties
of each response (rhythm, straightness etc) are automatic from each
light aid, the horse becomes relaxed in its body because the aid predicts
a precise response. The horse’s world is now predictable. The
horse’s general life is calmer too - things like separation anxiety
disappear as well as other nervous tendencies. This is because, unlike
before when responses were more random and less precise, the horse
is now able to ‘read’ humans. More than anything else it
is responsible for establishing what are known as rapport and trust
between horses and humans. The horse is no longer insecure and whinnying
for its mates shouting: “Help, I can’t read humans, I can
only read horses, is there anybody out there?” Think about this.
Does a small squeeze of the legs result in the right response but a
bigger kick result in the opposite response, i.e. slowing or showing ‘piggy’ behaviour?
Obedient, well trained horses are like obedient, well trained dogs – they
just don’t call out or become controlled by their environment.
They are ‘on the aids’. Of course young horses going out
for the first few times can be expected to be more nervous, but after
five or so outings they should become unfazed by new surroundings if
they are on the aids and their responses are consistent.
POSITIVE REINFORCEMENT
There are many associations that we train into the horse’s repertoire
of signals. One of the most important is the acquisition of verbal
praise such as ‘Good Boy’ for reward. Few people consider
how the horse acquires this as a positive reinforcer, and consequently
few horses adequately respond to it. We assume the horse knows what
we mean, as if the horse has some kind of pre-programmed English vocab
in its head. Because verbal commands have to be learned, they are called
secondary reinforcers. To be effectively learned they have to be associated
with a primary reinforcer, such as food or scratching/caressing the
horse at the base of the withers (a proven site where French researchers
showed lowered heart rate more than any other site). Because the base
of the withers is so close to the hands of the rider, scratching/caressing
is the most useful primary reinforcer. To train the horse to respond
to ‘Good Boy’, the words should occur just before and during
the scratching. The words and tone should be the same each time. Soon
the words come to evoke the relaxation that results from scratching
at the base of the withers. The words should be re-associated with
the wither scratching from time to time.
PRECISE TIMING
The light aid should be attached to the stronger aid that comes after
it. It should not be isolated by a gap in time before the stronger
aid. In horse training, the time gap between the light aid and the
stronger pressure should be the same as the time gap in between footfalls
in the rhythm of the gait - all responses should have occurred by the
count of 3. In other words the stronger pressure comes very swiftly
after the light aid. Training horses effectively therefore requires
skill and speed in decision making and action. In a fraction of a second
you have to decide if the horse responded adequately to the light aid,
and if not back it up with stronger pressure which is subsequently
released the instant the horse gives the correct response. This way
the horse optimally learns the light aid, and it rapidly evokes the
correct response without any increase in pressure of the aid. This
is the aim of all horse training – to transform the pressure-release
training rapidly to lighter versions of the pressure, and then later
to transform those aids into seat, weight and position aids. Through
careful repetitive training, the horse eventually ‘rote learns’ his
responses to the light aids so that he can avoid stronger pressures
altogether.
Next month I will describe the Exclusivity Principle, which is about
the importance of keeping aids and their responses unique - and what
happens when they are not…