In Defence of Dogs

by John Bradshaw

  Further observations of dogs’ interactions can provide more clues about how they manage their own behaviour when encountering other dogs. Each cycle of refinement of dogs’ ‘rules of thumb’ has a slightly smaller effect than the previous cycle; or, to put it another way, the first few encounters can be crucial in determining the form of a ‘rule’. For example, if the older dog had been unwell or in pain when the puppy first came into the house, it may have responded by snapping at the puppy when it came near. The puppy, in its next few interactions with the older dog, might then recall the tinge of fear it felt in those first few encounters. If by then the older dog was well, and able to behave in a more friendly way, that fear would soon be forgotten. However, transfer this scenario to a dog that the puppy meets only occasionally, perhaps a neighbour’s dog, and the fear may become a permanent fixture of the relationship. Not only that, but the puppy may generalize its fear to all other similar dogs: ‘medium-sized brown dogs with bushy tails make me scared’. This feeling may become very persistent if nothing happens to counteract it, emphasizing the need for owners to be careful when introducing a puppy or young dog to other dogs. It can also explain why two otherwise quite well-behaved dogs fight when they first meet, if each has previously developed a fear that is triggered by the appearance of the other. A Labrador that has previously been attacked by a small brown terrier will start to feel anxious whenever it sees a similar dog, and this will quickly become evident from its tense body posture. Moreover, if the terrier that has unwittingly triggered this anxiety has a history that makes it fearful of black dogs, and neither is in a position to back down (for example, they meet on a narrow path), then both may try to alleviate their fear by turning to anger and attacking the other.7

  Dogs that live in the same household can usually overcome such setbacks; they literally learn how to get along, in ways that the theory of ‘dominance’ overlooks. To observers, the relationships that arise can look like a set of ‘dominance relationships’, which taken together look like a ‘hierarchy’. In my study of French bulldogs, one of the bitches could be described as ‘dominant’ because the other bitches usually (but by no means always) deferred to her. But we have no evidence that the dogs themselves saw it this way. It is much more likely that the other three bitches in the group, one the daughter of the ‘dominant’ one and the others older, unrelated animals, each individually remembered her as being grumpy towards them when food appeared. Recalling this would not make them anxious, because their owner saw to it that they got enough to eat, so it would not – and, indeed, did not – cost them anything to give way.

  The ways that dogs interact with each other when they meet can thus be explained without reference either to ‘dominance’ in general, or to the captive wolf-pack model in particular. Even in groups of dogs that live together, what might appear to be a hierarchical structure is almost certainly a projection of our preconceived notions of canid relationships. There is no evidence that suggests that all dogs are motivated by the desire to achieve ‘status’ over other dogs; although some dogs undoubtedly appear more competitive than others, in all likelihood they are simply more strongly motivated to compete for things they happen to value highly – toys, for example – without being even slightly aware that they are thereby achieving something that we might (misleadingly) label ‘status’. Of course, we cannot be certain that wolves have this capacity either, so it is quite possible that their ‘hierarchies’ are also apparent only to us, and not to the wolves themselves. As predicted by the RHP model, each individual dog (or wolf) is likely to be using its experience of previous encounters to gauge how to behave, each time it interacts with another of its own kind; these ‘rules of thumb’ enable it to coexist comfortably with every other individual in its group. The range of experiences each individual can call upon will vary depending on how familiar it is with the other dog, but whether the other individual is totally unknown or one of its lifelong companions, the dog will use what information it has available to judge the other’s behaviour and proceed in the safest but most effective way possible. Consideration for its own safety will inhibit aggression, which is dangerous to both parties, while its experience of previous encounters will cause it to focus on gauging whether the other dog appears interested or uninterested in the resource in question. The majority of encounters between dogs pass without incident, with neither of the participants knowing or caring that its ‘status’, as some experts would have it, may have been affected.

  Unfortunately, the idea that dogs see everything in terms of ‘status’ has been embraced most emphatically in interpretations of relationships between dogs and their owners. In the popular conception of these relationships, the dog perceives the owner as just another member of the pack – and as an obstacle to the dog’s accumulation of ‘status’. By encouraging owners to believe that their dogs will, at some point, try to ‘dominate’ them and take control of the household, this idea promotes the use of ‘status reduction’ techniques, and physical punishment if these fail. But if the dog has no concept of ‘status’ – and we have reason to believe it does not – then none of these techniques is going to achieve precisely what is intended. Some techniques (especially the punishment) will alter the dog’s behaviour, but not necessarily in the desired direction.

  Many dog trainers still rely heavily on the idea that most dogs try to control the human families that adopt them. In this way of thinking, dogs must conceive of humans as members of their own species, albeit rather strange two-legged ones, and their behaviour towards people must therefore be derived from what occurs in wolf society; this has been termed the ‘lupomorph’ model.8 There may be some truth in this model; after all, if it were completely false, such that dogs think of humans as completely distinct from them, then it seems likely that dogs would have to have evolved a totally novel set of behavioural responses towards humans, at an early stage in domestication. But that conclusion seems implausible, given how much overlap there is between how dogs interact with humans and the ways they interact with each other. Still, some dog trainers have taken to the extreme the notion that dogs think of humans as being similar to themselves.

  The model adopted by many dog trainers has generally followed the outmoded view of dogs as constantly struggling to assert dominance, over their own kind and hence over humans as well.9 In order to account for the generally harmonious nature of most dog-owning households, these trainers assume that some dogs automatically ‘respect’ the superior status of their human pack members, possibly because they come to realize that humans are smarter than they are, or alternatively, simply because the humans evidently function like parents in providing food. However, they portray other dogs as attempting to achieve, or actually achieving, dominance over one or several family members and repeatedly exhibiting aggressive behaviour in order to affirm their position in the ‘social hierarchy’ of the ‘pack’. In this view, the undivided attention that dogs give to their owners and families is unwavering surveillance for an opportunity to move upwards in the hierarchy.

  Many dog trainers and behaviour experts still wholeheartedly support this concept – despite the fact that science has almost completely repudiated it – and have even come up with rules designed to thwart dogs’ supposed attempts at domination. According to these trainers, the ‘dominant dog’ constantly gives himself away by his body-language. If he puts his chin or paw on his owner’s knee, it means he thinks he is taking control of his owner’s behaviour, and is therefore on the road to becoming the pack leader. To forestall this attempt at ‘domination’, they go on to assert, owners should always move the dog’s paw or chin off their leg. However, for some reason such trainers make exceptions for small dogs: if these are used to sitting on their owners’ laps, this does not necessarily mean the dog thinks he is dominant. As an additional measure to prevent dogs from assuming dominance, owners are urged always to go through doors and gates in front of their dog.10 Some trainers have even come up with elaborate lists of these
‘commandments’ – such as those in the following example – that are supposed to stop your dog from thinking it is dominant over you.

  Do not allow your dog to eat its meal until you (the top dog) have eaten first.

  Do not allow your dog to leave the house (den) before you (the top dog) have passed through the doorway first.

  Do not allow your dog to climb onto the sofa or bed (only top dogs are allowed to rest in the cosiest places).

  Do not allow your dog to climb your stairs, or to peer at you from the top of the staircase.

  Do not allow your dog to peer into your eyes.

  Do not cuddle or stroke your dog.

  Do not interact with your dog unless you are involved in some kind of training.

  Do not greet your dog when you come home from work or from the shops etc.

  Do not greet your dog first thing in the morning; it should be the one to greet you (the top dog).

  Do not allow your dog to keep the toy at the end of a game; it will interpret this as winning.11

  The effects of these ‘commandments’ vary greatly, but none of them are especially constructive. Indeed, if dogs do not have a concept of ‘status’, and there is no evidence that they do, then some of these recommendations will be either harmless or incidentally beneficial to the dog–owner relationship. For example, many owners will prefer not to encourage their dogs to go upstairs, or sleep on their beds with them, though there is no evidence that allowing either would in itself have any effect on their relationship in general. Other ‘commandments’, however, such as the admonishment against cuddling or stroking the dog, seem aimed at taking much of the pleasure out of keeping a dog, turning dog-keeping from a joy into a challenge.

  Some of these ‘commandments’ have been investigated scientifically, but none of those examined were supported by the research. In one study by my colleague Dr Nicola Rooney, dogs were allowed to win tug-of-war games played with a person, over and over again; understandably, this made the dogs more keen to play with people than when they were forced to lose every time, but there were no signs indicating that any dog became ‘dominant’ as a result. In another of her studies, owners reporting that they always let their dogs win games were found to be no more likely to have disobedient dogs than owners who always insisted on winning, whereas dogs whose owners liked to play contact games, such as rough-and-tumble, were noticeably more attached to their owners than those that were usually kept at arm’s length. Not only were dogs that had been allowed to break the ten commandments listed above not in control of their owners’ behaviour, but they were also no more aggressive, which they should have been if their owners had inadvertently given them the green light to take over the household.12

  There is one more problem with the dominance theory of training, and it is especially significant: even if wolves themselves are inclined to dominate each other, it seems unlikely that domesticated dogs would have retained this desire to dominate. Even if we believe that dogs can only perceive us as if we were other dogs (or wolves), and even if we accept the dubious assertion that canids have a drive to dominate other canids, there is no logical basis for assuming that they would automatically want to control us. Domestication should have favoured exactly the opposite; dogs that passionately want us to control them. It seems very likely that, in the early stages of domestication, any dog that tried to take control of the human family it lived with would have been rejected in favour of one that was more biddable. Thus, even if there are some wolves that do have this hypothetical drive to dominate their ‘pack’, and this character trait is heritable, it seems highly unlikely that wolves with this trait would have contributed significantly to the ancestry of the domestic dog.

  It has become abundantly clear that the model upon which many people are training, managing and simply interacting with their dogs is fundamentally wrong. The traditional ‘lupomorph’ model is contradicted both by the current conception of how wolves actually organize their lives and by the logic of the domestication process – and since it does not explain dog–dog social behaviour adequately, it is also highly unlikely to be of any use in explaining relationships between dogs and their owners. Nevertheless, this model is still promulgated by many dog trainers, who use it to justify their methods. These trainers frequently portray owners as misinterpreting their dogs’ motivations, and promote the ‘dominance’ model as the only way to restore healthy relationships. At the very least, this approach saps the joy out of dog ownership; at worst, it is used to justify physical punishment as an unavoidable component of training. Many people now think that these punishment-reliant training methods are unnecessarily stressful for the dog. While they may appear superficially effective, such methods often do not work well in the long term, for reasons that are abundantly clear to those scientists who study how animals learn.

  4

  Sticks or Carrots? The Science of Dog Training

  Currently, dog training has a high profile in the media; evidently it makes for good television, as evidenced by the rise of celebrities like Cesar Millan, the ‘Dog Whisperer’, and Victoria Stilwell, presenter of It’s Me or the Dog. But there is tremendous disagreement among dog trainers about the best approach to shaping dog behaviour. A number of high-profile trainers and behaviourists continue to promote the idea that dogs are pack animals and that many can be controlled only through the application of ‘dominance’ theory and the use of physical punishment. For example, Cesar Millan writes: ‘Dogs have an ingrained pack mentality. If you’re not asserting leadership over your dog, your dog will try to compensate by showing dominant or unstable behaviour’.1 Or this from UK ‘Expert Dog Trainer and Canine Behaviourist’ Colin Tennant: ‘Most dogs will strive to dominate any other dogs or humans with whom they come into contact by body language and/or growling, biting or aggressive physical bullying.’2

  Others, such as Karen Prior, Patricia McConnell and Jean Donaldson, radically disagree with this approach, rejecting the wolf-pack analogy and advocating the training of dogs as if they were any other animal. Furthermore, they emphasize that training should be based around rewards and abhor the unnecessary use of physical punishment. Dr Ian Dunbar, one of the originators of this approach, states that compliance, the goal of all dog training, is most often achieved through positive training methods, specifically the lure-reward methods – using treats and praise – that he pioneered. Dunbar, a veterinarian and dog and puppy trainer with more than twenty-five years’ experience, bases his methods soundly on dog psychology, backed up by a doctorate in animal behaviour from the University of California, Berkeley, and a decade of research on communication and behaviour in domestic dogs. The debate between the two camps has at times become heated, even personal. For example, Dunbar has said of Millan: ‘He has nice dog skills, but from a scientific point of view, what he says is, well … different. Heaven forbid if anyone else tries his methods, because a lot of what he does is not without danger.’3

  These differences of opinion are not just of interest to the dog trainers involved; they have real effects on the welfare of dogs. Every year, many dogs are abandoned, even euthanized, because they have come to behave in unacceptable ways. In many cases, these behavioural problems are the result of inept or inconsistent training. For this reason, it is essential that we try to understand how dogs really learn, and thus which training methods and philosophies are most effective. Getting training right is essential both to the welfare of dogs and to the peace of mind of their owners.

  The notion that dogs are wolves under the skin still pervades much of dog training today, despite having been abandoned by the scientific and veterinary communities and an increasing number of dog trainers. The ‘wolf pack’ approach promotes two interdependent ideas: that dogs, because of their ingrained ‘pack mentality’, can be controlled only if their owners adopt the role of pack leader, and that the most reliable way to ensure this outcome is by the use of physical punishment. Both these ideas go back to at least the nineteenth century, b
ut were given added reinforcement, or so it seemed at the time, by the studies of captive wolf packs that were conducted in the middle of the twentieth century. Now that a more accurate picture of wolf society has emerged, based on family ties, the credibility of both ideas has been seriously undermined, yet they are still widely promoted.

  The widespread use of punishment-based dog training is usually traced back to Colonel Konrad Most, whose highly influential book Training Dogs: A Manual first appeared in 1910 (in German) and was translated into English, thanks to popular demand, in 1944. Most was emphatic that the relationship between man and dog was not only hierarchical – with only one ‘winner’ – but also could be established only through physical force, by an actual struggle in which the man was instantaneously victorious. The dog had to be convinced of the absolute physical superiority of the man.4 This approach demands that the owner constantly maintain and reinforce his or her position at the head of the family pack. In this conception the dog perceives the people it lives with as fellow members of its pack, and misbehaviour is construed as being due to a failure of the owners to maintain their dominance over the dog. Training methods are accordingly designed to lower the dog’s position in the hierarchy, or ‘pecking order’ as it is sometimes referred to.

  The Monks of New Skete, bestselling authors of dog-training manuals for more than thirty years, are highly influential promoters of this philosophy. They maintain that understanding wolf behaviour will help owners to understand their dogs, and that many books about wolves are often of more use to owners who wish to understand and appreciate their dog’s behaviour than dog-training manuals.5