Positive reinforcement works, and nothing that I say in this video is intended to suggest otherwise. What I am questioning is the increasingly common advice that if we want a dog to perform a behaviour faster, with greater enthusiasm or with more determination, all that is required is that we selectively reinforce the better examples of that behaviour and withhold reinforcement from those which fall below the standard we are looking for.

In theory that sounds perfectly reasonable. If faster responses are reinforced and slower responses are not, we would expect the faster responses to become more likely. Where I think we need to be considerably more careful is in taking that perfectly legitimate principle and extending it into the suggestion that we can continue the process indefinitely, progressively producing faster, stronger and more enthusiastic responses simply by continuing to raise the threshold at which reinforcement becomes available.

Dogs are not blank slates and they do not all arrive at training with the same genetic, physiological or behavioural characteristics. They differ enormously in their natural speed, motivation, temperament, persistence, previous learning and in what they find reinforcing in the first place. Of course training can improve the performance of an individual dog, sometimes considerably, but improving what is already there is not the same thing as demonstrating that we can manufacture whatever degree of speed, enthusiasm or motivation we happen to want simply through the selective delivery of reinforcement.

I use working dogs as an example because I think they demonstrate this rather well. People breeding, selecting and training dogs for police work, detection, military work, protection work or high-level competition don’t simply take whichever dog happens to be available and assume that, with sufficiently clever reinforcement, they can eventually turn it into the dog they require. Considerable time, effort and money goes into breeding and selecting dogs which already possess characteristics suited to the work, and even then some of those dogs will prove unsuitable as they mature. If selectively reinforcing increasingly good performances could reliably overcome those individual differences, much of that selection process would be unnecessary.

The problem becomes even greater when we leave controlled training conditions and enter the real world. The environment is full of things that we don’t control and which may be considerably more reinforcing to the dog than anything we happen to have in our pocket. The dog that would happily turn immediately for food, a ball or praise in one environment may make a very different choice when following a scent, seeing another dog, encountering wildlife or being presented with an opportunity that has considerably greater value to that individual at that particular moment.

So if you have been diligently rewarding the faster recalls, trying increasingly valuable food, becoming more animated, producing toys and doing everything you have been told should progressively increase the speed and enthusiasm of your dog’s response, yet you feel that you have reached a plateau, I don’t think it is reasonable simply to conclude that you must be doing something wrong. There may simply be a limit to what that particular reinforcer, in that particular dog and in those particular circumstances, is capable of achieving.

Towards the end of the video I also discuss negative reinforcement, because reinforcement is reinforcement regardless of whether it is classified as positive or negative. Both describe processes through which behaviour is strengthened, although they achieve that through different contingencies. Where positive reinforcement is insufficient to produce the degree of responsiveness or control required, particularly where safety is involved and the environment contains powerful competing motivations, I don’t believe that another form of reinforcement should be dismissed simply because the word used to describe it happens to be “negative”.

Ultimately, I am far more interested in what actually happens to the behaviour, what effect the training has upon the dog, and whether what we have taught continues to work when we leave the carefully managed training environment and take that dog back into the unpredictability of everyday life.

For more detailed discussions like this, practical training demonstrations, live Q&As and opportunities to submit your own training videos for analysis and feedback, you can join me in Jamie Penrith’s Canine Classroom.