One of the aims of this site is to promote positive reinforcement training in the field of dog training for law enforcement. Traditional methods have leaned too heavily on the dominance theory that wrongly advocates harsh treatment to maintain top dog status. It also relies too heavily on punishment and negative reinforcement in training. This article is taken from the book Positive Police Dogs: Philosophy which will be available here shortly. It summarises the reasons set out in the book as to why punishemnt doesn’t do what people think it does, and the negative side affects of relying on punishment in training dogs.You cannot teach anybody anything using punishment alone. It isn’t and shouldn’t be regarded as a teaching tool.
It may have an effect on future behaviour but as we have seen the outcome is unpredictable.
The unwanted side effects of punishment often work against our end result.
Stress has a cumulative effect and overuse of punishment can lead to high levels of stress which apart from being a block to learning can lead to other undesirable behaviour.
Relying on punishment is often frustrating for the trainer and escalation of the punishment is the natural remedy if things don’t improve. Dishing out harsh punishments alters your mood in a negative way and the risk of becoming emotionally out of control is real. It harbours resentment and can lead to very punitive behaviour in order to off load the growing anger towards the dog as things (inevitably) fail to improve.
Whilst addressing one issue with punishment you are often punishing another, desired behaviour.
The ultimate power of punishment comes from evoking a fear response from the dog. The only way it will have any lasting effect is if you can create such a sense of fear in your dog that he wouldn’t dare do it again. We have alrerady seen that they often just learn to do it in your absence. Either way, is making your dog that fearful of you or of doing some act conducive to creating a strong, confident Police dog?
Skinner stated that positive reinforcement was the best influence on behaviour and for training dogs where the confidence of the trainee is important it is hard to argue otherwise. Although Campbell and Church (1969) found that punishment could have a strong effect on behaviour they acknowledged the side effects of anxiety, stress, withdrawal and aggression in the subjects.
Estes (1970) work confirmed the findings of Bow and Church (1967) and found that all punishment did was surpress behaviour. The strength and duration of the supression effect depends on the intensity of the punishment but the behaviour is merely supressed not unlearned.
Estes work showed that animals that were punished showed an initial halt to a previously learned behaviour but that the behaviour soon reappeared and it took as many unrewarded performances of the behaviour to become extinct as for those animals that weren’t punished.
So if you rely on punishment you are just delaying the problem. The behaviour you wish to address will resurface because you haven’t addressed the motivation of the dog. You, as we have already seen, will have been positively reinforced for using punishment because you interpreted the surpression as success. The cycle has been set in motion.
Punishment should be saved for dire emergencies when stopping the behaviour is all that is required. Training for every eventuality should avoid the need and should the need arise then your conditioned aversive should be effective. If punishment is ever used it should be a time to reflect on your deployment of the dog and training to see which was at fault and then learn from it.