Don't Shoot the Dog: The New Art of Teaching and Training


 Click on the cover above to go to this book at

 ‘Don’t Shoot the Dog’ has become a classic text in dog training, and was the first popularisation of operant conditioning for dog trainers and owners, though others had previously used this technique to train a range of animals, notably BF Skinner. The idea of operant conditioning had been around for decades before ‘Don’t Shoot the Dog’ appeared in 1984, but this book was a key landmark in history of dog training. Karen Pryor is also usually associated with clicker training, which became popular among many trainers and owners, especially at competition level, during the 1990s. Why is her book so popular, and what are its strengths and weaknesses?

The reason this book is so popular probably lies in its offer of a relatively easy way to train just about any living creature, including children and husbands, not just dogs. It also fits in with a trend towards using rewards, whereas previously there was more usage of punishment in training. (Pryor herself distinguishes between positive reinforcement and rewards, but many trainers use these terms interchangeably, and I will call her style reward-based training for reasons later explained.)

This trend towards reward-based training has not only affected dog training in Western Europe, the US, and Australia, it is part of a wider trend in the way humans have come to treat children and other creatures that are dependent on us. ‘Spare the rod and spoil the child’ was still heard in the UK in the 1950s, when I grew up, but today there is a wide consensus that children should not be beaten. Beating children and dogs out of temper was seen as a ‘bad thing’ in those days, but some physical chastisement was often seen as necessary for their proper development. This move away from physical punishment does not, however, mean that all punishment is a practice of the past, since there are more ways of punishing others than beating them. Ignoring children and dogs can be a form of punishment.

Perhaps the move away from the more physical forms of punishment is because we have become ‘less physical’ in our everyday lives, for example, more of us do ‘mental’ rather than manual labour, and we use computers to communicate with people in other countries, rather than greeting friends with handshakes or hugs. It may be that some of the move away from punishment is illusory, since, though dogs and children are beaten less in the UK these days, other forms of punishment are often not noticed by those inflicting them.

Strengths of Karen Pryor’s book include clear explanations of why rewards can be more effective than punishment, and of the need to reinforce the behaviour when it is happening. Clickers are useful for this reason, because they give the dog a signal that what he is doing is right. You can click faster than you can get out a titbit, and the click doesn’t interrupt the dog’s activities in the way that stopping to feed a titbit does.

The emphasis on variable schedules is also useful, since a common mistake of owners is to reward every desired action with titbits. This can leave owners feeling helpless when their pockets are empty, and fixed schedules are less effective than variable schedules. Concepts such as shaping and conditioned responses (eg using clickers) are also clearly explained. The chapter on ‘Untraining’ is especially interesting and helpful, in that several different approaches are presented, allowing readers to make comparisons and think these issues out for themselves.

Some people are extremely enthusiastic about this book, and even go so far as to say that it is the only book that owners need to read. I would, however, disagree, for a number of reasons.

Firstly, this is not a book for novices. It is not a ‘how to’ book, rather it goes into positive reinforcement and operant conditioning in general in abstract terms with many examples from outside dog training, leaving readers to work out how to apply it to their own dogs. Novices with a new pup or a rescue dog are likely to need a little more help.

Secondly, many people may find the reference to dog training in the title confusing, since Pryor often seems to mention just about any animal other than dogs. In many ways, her work would have been more useful if she had stuck to relations between humans and dogs (or other non-humans), and not tried to extrapolate her ideas to people, presenting operant conditioning as a miracle cure for dealings between humans. Yes, I agree that sometimes people forget common sense in their dealings with one another, for example, getting angry with partners who return home late. (Like dogs, humans tend to return home faster when they have a nice welcome.) However, our relationships with other humans can go beyond what operant conditioning has to offer. Cognitive psychology is usually seen as a more effective way of getting people to change their own and others’ behaviour, according to my friends in counselling. I’d have liked to see her try to mesh some of her insights from behavioural psychology with those offered by cognitive psychology, because, though behaviourism can offer useful insights, it is a bit limiting. Perhaps this is asking too much of one book, but Pryor does seem a little too messianic in her zeal to preach operant conditioning as a cure-all, so I think she lays herself open to such criticisms.

We can’t reason with dogs, and here behaviourism has more to offer as a direct and simple way to get through to them. However there is another gap that worries me when reading Pryor and other authors of ‘dog books’ who stress operant conditioning, and that is their tending to neglect or downplay the influence of inherited traits, including breed characteristics. Dogs may also inherit predispositions to be nervous, ‘pushy’ or whatever, that are unrelated to breed. And dogs are hardwired to do certain things such as jump up to greet people, because that is the nature of dogs. A novice might take Pryor’s text and think of dogs as a blank sheet, and see all behavioural problems as a result of inadequate training, or the dog having had some past experience of being rewarded for a certain (‘bad’) behaviour, or being frightened, so being ‘bad’ out of fear. Yet some breeds have inbuilt tendencies, eg Keeshonds tend to bark a lot. A Keeshond barks because it’s in the nature of the beast. Operant conditioning can help to some extent in controlling the barking, but it won’t turn a Keesie into a quiet breed.

‘Don’t Shoot the Dog’ is not really a good text for learning about operant conditioning either, I certainly found the explanations of negative reinforcement confusing, and prefer someone more analytical, like Mary Burch, who sets out the theory very clearly in ‘How Dogs Learn’. There’s more repetition, more of a polemic, and less of a clear-cut explanation in ‘Don’t Shoot the Dog’. Pryor stresses that positive reinforcement is not the same as giving rewards, but is more effective. However, since the term includes both signs that a dog is about to be rewarded, and the actual reward, I have no problem with calling her style ‘reward-based training’. Her account would have been easier to understand if she had simply said that reinforcement involves stimuli that make behaviour more likely to happen, whereas punishments make it less likely to happen.

I have found symbols and a schematic presentation helpful in following the logic behind operant conditioning. The R+ and R- and P+ and P- symbols are especially useful. All the plus means is that something is added (positive), whereas a minus means that something is taken away. The R stands for reinforcer, and the P for punisher. So:

R+ = positive reinforcement = a reward or promise of a reward that makes a dog more likely to do something (eg titbit for a food-oriented dog),

R- = negative reinforcement, removing something nasty to make it more likely that a dog will do something (eg slackening a taut leash to help a dog socialize better),

P+ = positive punishment = punishment that makes a dog less likely to do something (eg shocking dog to curb chasing), and

P- = negative punishment = taking something nice away to make a dog less likely to do something (eg owner turning away from jumping up dog, where owners’ attention = nice for the dog).

This is what I missed in Karen Pryor’s explanation of behaviourism. People learn in different ways and perhaps other readers might have less trouble following her, but unless I have these concepts set out in a fairly schematic way, I find them difficult to follow.

There is another criticism that could be made of this book, which is perhaps a little unfair, but I’ll make it anyhow…and that is that it’s fairly ethnocentric. The sorts of things we ask of our dogs vary over time and from place to place. We have more companion dogs, and fewer working dogs, more pedigrees, and fewer mutts. OK, so most dog books sold in the UK tend to be ethnocentric, why lambast Karen Pryor? In my case, a) because I happen to be reviewing her book, and b) because I have lived in rural Spain, and have seen dogs brought up and kept in very different ways from in the UK. They tend to be either working dogs, purpose-bred for guarding and hunting, and mostly relatively isolated from humans, or village mutts, a kind of nondescript smallish type of dog. These dogs trot round villages minding their own business and bothering no-one. They tend to avoid people they don’t know, and cars. They are also well-behaved by UK standards. Why? They don’t usually get formal training. Maybe the ones that don’t behave well die young - from the kick of a passer by, or being run over. Dogs that venture into other people’s stockyards often die of poisoning. So to some extent, bad behaviour is bred out of them.

This link between cultural change and the type of breeds we keep, ties in with a point made forcefully by the Coppingers in ‘Dogs’, which is that many dog behavioural problems seen in western societies are the result of people keeping breeds in ways that they are not suited, keeping herding dogs like collies shut up and bored all day, for example. A dose of operant conditioning for ten minutes in the evening is not going to be enough for a border collie kept shut up in a house all day. It’s often worth looking at our own culture through the eyes of another culture to see why we sometimes get things wrong. I’m not proposing keeping small Spanish village dogs and allowing them to run loose. They would be mowed down fast in urban UK traffic, and tend to have pretty short lives in rural Spain. I am saying, however, that some of the problems that Karen Pryor is trying to address have their roots in cultural change, especially fashions for keeping very specialised working dogs for companionship. This fashion has also started to extend to rural Spain - by the 1990s there was a fad for keeping Siberian huskies in the villages in central Spain I know well. Not the best dog for a region that is hot even by Madrid standards.

In conclusion, Karen Pryor’s book is a ‘must read’ for anyone seriously interested in dog behaviour and training, but it should be supplemented by other texts. It is not the first book I’d recommend for a novice, because it doesn’t have enough in the way of detailed instructions. Ian Dunbar’s ‘How to Teach a New Dog Old Tricks’ is perhaps the best book for a novice. ‘Don’t Shoot the Dog’ is not analytical enough for someone who is studying animal behaviour in an academic way. It could also lead people to have unrealistic expectations of their dog, if they have chosen a breed that is not suited to the lifestyle they lead. So why bother with it? Firstly because it’s an important historical document, secondly because it’s a good read, and thirdly because, despite its shortcomings, ‘Don’t Shoot the Dog’ is very clear on the rationale behind some of the reasons why certain training techniques tend to be more useful than others.

Mary Burch and Jon Bailey (1999) ‘How Dogs Learn’
Raymond and Lorna Coppinger (2001) ‘Dogs’
Karen Pryor (1999) ‘Don’t Shoot the Dog’ (revised edition)

A.L. April 2002