Book: Teaching with Reinforcement 


Author: Kay Laurence
Punishing a behaviour is quite separate from punishing the dog.
A punisher is just the term used to measure the outcome. A behaviour has been “punished” because the behaviour diminishes, fades or happens less frequently. Very often we will simply remove the reinforcer, or opportunity for reinforcement and encourage the behaviour to fade in its own time.

For positive trainers there is a similar misuse of the term “punish” as there is for “reward”, we get stuck on the intent of the deliverer, and not measuring the outcome of either on the behaviour. As soon as our head, and our heart, is around that distinction we can become very effective in using reinforcement and punishment without ever diminishing the dog, or our relationship.

Jumping up at people to seek approval is a juvenile behaviour and a perfectly normal behaviour for young dogs. If the pup is maintained at floor level through the approval process and NEVER reinforced for elevating, then the behaviour of trying to jump for attention will not be reinforced, and not develop. Dogs are so much more skilled than people at becoming aware of the outcome of a single event. You have to be on your toes through all interactions with young animals, and ruthless in managing the behaviour of other people. One careless visitor, who “doesn’t mind” the puppy jumping up at them (I wonder about why that is reinforcing huh?), can make a lifetime of bad habits. In my experience when an instinctive behaviour is allowed to be reinforced on just one occasion, the dog will continue to attempt success in that behaviour many hundreds of times. On the other hand, if a learned behaviour is only successful once, and the success is not repeated many hundreds of times, it usually fades instantly. The key to the fading rate relative to the reinforcement schedule is the instinct vs. learned behaviour. Using that instinctive behaviour does not need external reinforcement.

I have access to exercise my dogs on stock free farm land. Quite often deer cross our path, and neither of my breeds are fast running chasers. The young collies spy their first galloping deer and take off, returning a few minutes later, tongue lolling out. An unsuccessful instinctive response, the deer can outstrip both Collie and Gordon with ease. Next occasion, the youngster doesn’t even take breathe to chase when they see a deer – the behaviour has extinguished, instantly. This is especially effective on the first occasion of the behaviour. Lack of success was a punisher, and for the collies they learn the same reaction with hare and pheasant. When they notice the deer they are inevitably busy using their herding instinct on their peers. This is a highly, and continuously reinforcing activity which is the norm for our exercise. To break from this behaviour and react to the deer, it would be a double value punisher. Firstly because they are not fast enough to catch the deer, and secondly because they ended a reinforcing behaviour – herding their peers. Perhaps if we were on a more relaxed walk, where there is very low value reinforcer for the collie – just browsing the hedges, deer chasing would be more reinforcing in comparison.

The Gordons have no reaction to deer or hare, but live to point the birds. Pheasant and partridge sit tight when they hear 85 pounds of predator crashing through the undergrowth. Without excellent scenting skills they would miss the birds but the many generations of selective breeding has fine-tuned the discriminating skills and they freeze as soon as they get the scent of the bird. The freeze is instantly transferred to any Gordon within visual reach. This is referred to as “backing”,  and dogs are assessed for this essential skill in competitive trials. Without the instinct to back, No.2 dog could carelessly crash into the birds if the wind was not carrying the scent to their location. The backing dog usually faces the pointing dog, not the location of the bird unless they also have the scent. The pointing dog now needs to hold the point until responding to the cue to flush. My dogs are not trained for shooting, so they respond as they choose on point. Kent over-freezes, and will stand quite still for several minutes. Mabel will very slowly, catlike, move one leg at a time to approach the prey – with every intention of going for a kill.

Interestingly the collies never respond to the Gordon’s point and vice versa. Both breeds back each other in stalking behaviour, but it does not cross the breeds. The collies usually serve the role of pheasant crashing to break Kent from his freeze.

Types of punishers

Most of the punishers we employ will be the removal of the reinforcer. Tessie is fixated on balls, especially those that squeak. At times this serves a useful purpose, where she “points”, or fixates, on the ball for a couple of hours, just shifting position every few minutes. Obviously a highly reinforcing behaviour needing no external reinforcement, she does not need anyone to interact with the ball to increase the pleasure. Occasionally a fast collie will steal the ball, and it usually takes her a few minutes to realise what has happened, she does not chase the collie with the ball or fixate or point to it. But if the ball enters the house the behaviour goes into high arousal, and the this needs to be managed. We have a successful (from my viewpoint) technique of removing the ball “to go to bed”, where it is placed in a basket in the cupboard. The fixating behaviour has been “punished”, I need to seal the ball inside a plastic bag, or the scent of the ball through the edge of the cupboard door will trigger fixation to the cupboard. In this behaviour the cue and the reinforcer are the same.

I also use this cue/reinforcer to avoid another behaviour. When out on a walk Tessie will harass Mabel, who is now too old for these politics, so the ball comes along for the walk, and relieves the stress for Mabel, keeping Tessie fully employed. Managing this reinforcer, manages the occurrence of the behaviour.

Very, very often the punishers and reinforcers are two sides of the same coin. You can withdraw attention and give attention, one will punish the behaviour, one will reinforce it. It is extremely difficult to be neutral. And like reinforcers, punishers have different values, can be continuous, and are hard to ensure the attach to single behaviours.

By transferring the reinforcers to the desired behaviour, we have “punished” the undesired behaviour and it will happen less frequently.


Submit a Comment

Your email address will not be published. Required fields are marked *

Get in touch

6 + 6 =

News on courses, articles and stuff you don't want to miss.