Praise and Punishment
The building blocks
Most animals and humans respond better to positive situations than
negative ones. Golden Retrievers being a breed that live to please their
owners are more likely to flourish under a training program that is geared
towards positive reinforcement than punishment.
Background
When we are training our dog to perform a particular new task, we are
effectively asking it to have a go at the unknown. For example: the very
first time we ask a dog to sit, it has absolutely no idea what we require
of it. It has two options: do nothing, or do something. The secret to dog
training is for the dog never to be afraid of doing "something".
As long as the dog makes an attempt to do "something", we can be
assured that sooner or later it will hit on the first step towards the
correct behaviour.... which we then have the opportunity to praise/reward.
But if the dog is afraid to do "something" and opts for
"nothing" the job is almost impossible.
The pleasures positive reinforcement
Using the positive reinforcement approach, the dog comes to believe
that every action is a valid one, but at times some are more rewarding
than others. Assume that I have taught my dog to sit and drop (lie down).
If I tell him to drop he can perform either of these actions. However he
knows that only one of them will bring the reward that he's after, the
other action is not wrong as such.... But it doesn't bring reward at this
time (it might later under different circumstances). So the dog chooses
to perform the action that is most rewarding to him, which also
happens to be the one I want! The dog that has had this sort of
upbringing/training is not the sort that collapses under pressure. Dogs
(and people!) that have trouble performing under pressure generally are
scared of making a mistake. Dogs trained with positive reinforcement have
nothing to fear from making a mistake.... They just don't get rewarded. So
they are striving to achieve well for their own benefit (which you have
groomed to match your goals), not to try and avoid punishment.
The problem with punishment
Punishment often gets out of proportion to the crime. People tend to
get emotional with a dog that doesn't perform as expected, especially if
there is an audience to exacerbate their embarrassment. The tendency to
"tell the dog off' is used more as a vent for the owners frustration
than for any benefit of the dog and we've all done it. The dog that is
persistently punished is learning through fear and a fearful environment
is not conducive to learning. Such dogs might learn not to perform
specific actions (ie not to steal rubbish from a bin), but it is very
difficult to teach such dogs to perform actions where they need to use
some initiative and creativity. For instance the Utility Obedience
exercise of running out to the box and sitting in it is harder to teach
such a dog. They don't have the confidence to try "something"
and see what happens, they are always expecting that nasty punishment to
fall from above.
The reward
The reward can be anything that tickles the dogs fancy: verbal praise,
a pat, food, a toy, the opportunity to fetch something (Goldens love that
one!). Use your imagination and try a number of items to find what the dog
likes best. You can also vary the reward to help keep the dog interested.
So how do you get the message across that
you don't want "that' 'action?
The key point is that we need to let the dog know that a particular
action is not the desired one (at this point in time) and encourage him to
try a different one. Behaviours are not "bad" they are just
unwanted at this time.
See each "failure" as an experiment, each one brings you
closer to success. So we need to find a way to tell the dog that this
action at this time is not the one we want.
The "try again '.'Word
Now that we have a whole new perspective on "bad" behaviour
we should use a different word to make sure we don't fall back into our
old habits. The problem with "no" is that it has so many
decidedly negative applications in life that we tend to over-emotionalise
its use. It's hard to keep your tone neutral and say "no" when
the dog has performed an action that we aren't after. It tends to become
"Nooooooooooooooooo" and becomes a word that causes fear in the
dog. Remember we don't want to scare the dog, just make it aware that
we're after a different behaviour. The word you choose isn't really
important, as long as it isn't a word you use very often (or the dog will
hear it all the time when it doesn't relate to him), and it Isn't
"no"! I use "wrong". It's an easy word to use in a
neutral tone, and is not in common use. I know someone who says
"that's nice, but what else can you do?". Note that it's almost
impossible say that with a harsh tone, or even a slightly disappointing
one.
Putting it together
With the dog paying attention and the reward handy just watch the dog
and see what behaviours it offers. If it just stops dead and won't do
anything, walk around a bit and try to get it moving. If, for example, you
want the dog to lie down you are waiting for any sort of knee-bend or head
lowering action. When it occurs, praise the dog and offer the reward. Note
that the first few times the dog gets rewarded it probably has no idea
what it did to "score". Be patient, you need to get the first
step happening regularly before you increase the criteria. For instance
you might be happy with a head lowering at first, but when it happens
regularly you might want a head lowering and a knee bend. When
that's regular, maybe you want a definite bow etc. "That's fine as
far as praise/reward goes, but what about the "wrong"
part?" I can hear you ask. Let's assume the dog barks to get the
reward and we want it to be quiet. You'd probably temporarily suspend the
original activity (getting the dog to drop) and concentrate on teaching
"quiet". Whenever the dog barks, you say "Wrong.......
Quiet" in a quiet tone (no point getting the dog all fired up with an
enthusiastic command, he needs to settle a bit) and otherwise ignore him.
If he's quiet for a second, reward him as you repeat "Quiet, good
quiet". Tell him "Quiet" again and try to extend the time
the dog will remain quiet. Each time the dog barks repeat "Wrong...
quiet", this lets him know that what he did last is not required
right now (although it might be later) what you want now is
"Quiet". Remember not to get all worked up about giving the
command, you should sound almost detached, as if reading the shopping
list!
The end result
If you are able to raise a pup under the principles listed above, you
should have a responsive dog that is not afraid to try new things. It will
learn quickly and be eager to please. Remember that no dog is perfect....
As no person is perfect. If the dog should misbehave the important
thing is to teach the dog not to repeat the behaviour, not to
"punish" the dog for the offence. Punishment teaches nothing,
except to fear the punisher. Remember also that a pup trained with these
techniques may also seem generally more "naughty" than those
trained under more standard "punishment type" methods. This is
because he's not afraid to try new activities, whereas the other dog may
be too inhibited to do much at all. Don't be annoyed or frustrated with
your puppy if this is the case, rejoice in the fact that he has an
outgoing personality and is easy to train.
What about my older dog?
The above techniques can be used on older dogs that have been trained
in a different manner, or not trained at all. Results will not be as
effective (or as quick) as they are with a pup, but any improvement is
worthwhile. An older dog will be less likely to offer a variety of
responses due to inhibition caused be previous training methods. A dog
that dislikes the whole process will be difficult to convert ... but not
impossible. Be patient with an older dog, they need more time and
understanding.
Further information?
The techniques outlined above are extended by the use of a
"clicker" as a secondary reinforcer. There is not a great deal
available at the moment in the way of clubs or books utilising such
methods.
Sherbrooke Obedience Dog Club runs classes using food/clickers.
A book called "Don't shoot the dog" by Karen Pryor discusses
the theory of the techniques and the principles that they are based on.
David Weston’s' book "Dog Training: The Gentle Modem
Method" shows how to teach a number of exercises using positive
principles.
[Back] [Next]
