Despite the dog training world slowly becoming more enlightened as the full benefits of positive reinforcement spread across the sector, it seems that there are still some out there promoting positive punishment as a viable training method. Newcomers are being taught methods that belong firmly in the last century despite all the information and research that points to more effective and humane methods. So why do the remnants of these outdated methods still linger on in some corners of dog training?
The first point to make clear is that positive punishment can work but (and it is a very big BUT) it has severe limitations and consequences which those who have an ounce of compassion and patience will regard as unacceptable.
The first problem is that positive punishment will result in what I call ‘learning in context’, in other words the learnt behaviour will only be exhibited in the context it was taught in, while the threat of punishment exists. An example would be a dog that is taught not to jump up by smacking him on the nose or kneeing him in the chest, he will stop jumping up to the trainer but will not generalise that behaviour to other people who do not deliver the same negative experience. It is no different with people driving, the whole driver safety education system revolves around punishment and the threat of punishment, take speeding for example. Speed cameras pose a threat of punishment and drivers ensure that they slow down when the threat is perceived but as soon as the threat is removed they speed up again.
The next issue is that of fear, anyone with a basic grasp of classical conditioning will understand that the pairing of an unconditional stimulus (pain) with a neutral stimulus (training venue) will lead to the training venue becoming the conditioned stimulus. In short, the dog predicts that entering the training venue will lead to pain and therefore elicits a fear response before any punishment has even been hinted at. No different than dogs becoming fearful of going to the veterinary surgery. It is also not beyond the bounds of possibility that something in the training venue (floor covering, noise, smells etc) will be generalised to other locations and the dog will unexpectedly become fearful for no apparent reason. The human equivalent would be fear of dentists, needles or even a dislike of a food or drink that made you ill. While fear can be a powerful component of classical conditioning it should have no part in training using operant conditioning, a process that drives the dog to work to achieve the desired behaviour rather than working to avoid punishment.
The other dimension of fear and other unpleasant stimuli is how they affect the capacity to learn. I am sure we all had a particular subject at school that we did not like, perhaps a teacher that was unkind or had an aggressive manner, inevitably this was a subject that we did not excel in, partly due to a lack of motivation but more importantly the readiness to learn is impaired by the state of stress during those lessons. Exactly the same principle applies to dogs, if it is fun there is motivation to perform, if it unpleasant the only motivation is to get it over as soon as possible or even disengage.
Some trainers will hide behind negative reinforcement in the belief that this clears them of inflicting positive punishment on the dogs. What often goes unmentioned however is that removing a negative stimulus requires a negative stimulus to be applied first, thus inflicting positive punishment before negative reinforcement can be exercised. In the case of a choke chain for example, the pressure of the collar on the throat cannot be released until it has first been applied and before the tightening collar is blamed on the dog pulling, take a look at who put the device on the dog. When negative reinforcement was first thought of by psychologists it was a process to describe how people might take an aspirin to remove a headache, or move into the shade on a hot day (as will dogs) not as a description of a useful training method.
Given the disadvantages of using positive punishment as a training method you might think that would be enough to avoid it but there are also good reasons why positive reinforcement should always be the method of preference. Firstly it creates a stronger bond between trainer and dog, secondly it cannot, by definition, include any unpleasant or cruel stimuli, lessons learned are readily generalised to other contexts, the learning will be more long-lived and the dog will ‘enjoy’ the process. The only possible downside (that is hardly worth mentioning) is that it may, in some circumstances, take a little longer for the learning to become established.
So why do people still use punishment as a matter of course? Possibly because the results (including the limitations and consequences) can possibly be achieved faster than by more humane means, possibly because the trainers do not fully understand the subject of learning, perhaps it is because they are told it is OK by people who are perceived by the uninformed as being something of an authority or perhaps they are unable to accept that what they have been doing for years is plain wrong. Whatever the reason, if they do not have the patience and knowledge to deliver training in a humane and informed manner that will enhance the handler/animal bond and be an enjoyable experience for both, they should perhaps consider the possibility that they have made an unsuitable career choice.
It is not a great leap from inflicting pain to abuse.
Whilst I’d agree that positive reinforcement is the way to go, your argument is fundamentally flawed. Positive reinforcement is equally ‘learning in context’. My dog has only ever been trained by positive reinforcement. I follow this approach for the simple reason of not wanting my dog to fear me, which should be enough for any responsible dog owner. It cannot be denied however that following the positive reinforcement approach my dog, and the others in the class I attended behaved differently at the class than they did outside of it, generally far more obediently, as they associated the yard with the rewards for certain behaviours. Likewise it’s clear that dogs respond to different people differently, even if trained positively, as they associate treats with one person and not another.
The strongest argument to my mind against punishment is that dogs tend to be very sensitive to their owner, as any owner can see if they suddenly start jumping around the dog will join in, if they praise the dog it’ll show happiness. Likewise a stern “No’ can have an equally strong reaction in a dog if used sparingly and strongly, and is in many ways a ‘punishment’. There is no need to strike or otherwise physically punish a dog where strong words are more than enough of a deterrent.
Firstly I commend your approach to dog training and although you are right to say that positive reward can lead to learning in context the point is that it is very easily generalised to other contexts by continuing the exercises taught at the class elsewhere. Using positive punishment is far more likely to require re-teaching for each new context making for a miserable experience. I also agree that saying ‘No’ is far preferable to resorting to physical means but it can become over used and be a confusing stimulus for the dog as it tends to be used to attempt to stop a wide range of different behaviours. Teaching different commands for different behaviours such as ‘drop’, ‘down’, ‘leave it’, ‘quiet’ etc makes for clearer communication with the dog thus, more training opportunities, more success for the dog, better bonding and all the other associated advantages.