|
It is worth getting
to grips with some of the terminology here.
Although a little confusing at first, it is
the 'language of learning' and is understood by
trainers from all walks of life all over the world.
If we use this 'standardised' vocabulary when we are
learning about animal training, it really
helps to make sure we are all 'on the same page'.
Avoiding a lot of unnecessary confusion.
A scientifically proven learning process.
Operant conditioning is a well-studied process by
which learning takes place in many animals including
human beings. The principles have been known and
understood for many years now and if correctly
applied produce reliable, repeatable results with a
wide variety of species including ourselves!
Operant conditioning explains how punishment and
reinforcement interact to alter an animals behaviour
and to enable it to learn from and make use of the
natural environment in which it lives. Operant
conditioning is described in scientific terms and
the terminology used can be confusing to begin
with. Words like punishment and positive, have a
very different meaning from that with which we are
familiar. All is explained below.
Punishment. In the science of learning
theory, anything which makes the behaviour preceding
it less likely to occur in the future, is defined
as a ‘punisher’ and the application of the punisher
is defined as a punishment. The punishment does
not have to be painful or violent, or harsh at all,
for it’s application to be defined as a
punishment. All that is required is that it
makes the behaviour preceding it less likely to
re-occur. This means that the animal receiving the
punishment must find it aversive for it to be a
punishment. Understanding this is important both
with animals and people. A punishment is not
defined by how harshly we regard it, but by whether
or not the subject we are imposing it on finds it
aversive and to what degree. What is a
‘punishment’ for one dog, may not be a ‘punishment’
for another. Nor is a punishment defined by the
intention of the person supplying it. Letting a
firework off may not be intended to be a
‘punishment’ by our neighbour, but if it frightens
our puppy whilst he is carrying a ball for example,
strictly speaking the puppy has been punished for
carrying the ball and he will be less likely to
repeat the behaviour in the future. When we
refer to punishment in the context of training, it
is this definition which we are using.
Reinforcement
is the other side of the coin. Anything
which makes the behaviour preceding it more
likely to occur in the future, is defined as a
reinforcer, and the application of the reinforcer
is defined as ‘reinforcement’. This is not the
same way in which many trainers use the word
reinforcement when it is often used simply to mean
‘going over’ training until it is thoroughly
understood. In layman's language, we often
describe reinforcers as ‘rewards’ . A reinforcer
can be anything the dog finds pleasurable, it does
not need to be food, it can be toys, games, a
retrieve and so on.
Two
categories of punishment and reinforcement.
There are two distinct ways in which to create both
punishers and reinforcers. These are described by
the words ‘positive’ and ‘negative’ This is
confusing to begin with as many of us use these
words to mean good and bad. However, here they are
used in the mathematical sense of adding or
subtracting something.
So, a positive punishment is where something (which
the dog finds aversive) has been added to the
dog’s environment whether it be a gesture, a word,
a glance, or a slap. In correct training
terminology, these are all punishments. A
negative punishment is where something pleasant to
the dog has been subtracted or ‘taken away’.
In exactly the same way, a positive reinforcement
is where something pleasant has been added to the
dog’s environment - some kind of reward, food for
example. A negative reinforcement is where
something unpleasant has been taken away.
Negative reinforcement is rarely used in UK gundog
training but is common in the US and forms the basis
of the 'force-fetch' programme.
Negative reinforcement is controversial because in
order to remove the unpleasant stimulus (pain for
example) the handler must first apply it to the dog.
The Learning Quadrant
These two types of
consequence with their two categories of application
give us four outcomes to any behaviour and are often
described as the learning quadrant, and depicted
like this: |
