Selective breeding for appearance is largely a product of the last 100 years; for far longer – probably since the earliest stages of domestication – man has been breeding dogs for behaviour. This tendency has continued right up to the present day, as dogs’ roles continue to become ever more specialized and demanding. For example, the Guide Dogs for the Blind Association in the United Kingdom has developed a strain of golden retriever/Labrador retriever crosses that are particularly suited for guide-dog training. Much of this breeding has fitted dogs to particular working roles, thereby doing a great deal to strengthen the bond between man and dog. But in the contemporary West, where dogs’ working roles have diminished, some of the more exaggerated behavioural traits – such as indiscriminate chasing and forceful territoriality – can be unhelpful for dogs whose primary role is to be companions.
One clear ‘personality’ difference between breeds lies in the extent to which they enact the predatory behaviour of their canid ancestors. Some dogs show no predatory behaviour even when you would expect them to do so: wolves, on seeing a small animal running away from them, would instinctively give chase – as would many dogs, especially the hunting breeds. Others, especially some of the guarding breeds, seem oddly uninterested. Although training plays a part, these differences between breeds are, at their heart, genetic.
The sheep-guarding breeds are extreme examples of this kind of unresponsiveness. Originating around the Mediterranean, these include the Pyrenean mountain dog, the Italian maremma, the Hungarian kuvasz, and the Turkish karabash and akbash. Many of them are entirely or largely white, bred to look more like sheep and less like wolves. Traditionally, these dogs were raised with livestock and then kept with the flock to guard it against predators. They therefore treat members of the flock as part of their own social group, confining their aggression to whatever they perceive as threatening to themselves or their flock. They do attack and kill rabbits, so some of their predator behaviour must remain intact. However, they have been reported as not knowing what to do with their kills – they simply carry their dead prey around until it falls to pieces. All dogs may be born with the ability to perform the various elements involved in predatory behaviour, but in some breeds some of these elements do not appear until adolescence, and thus never become integrated into the rest of their behaviour.8
On the other end of the spectrum are working collies, that display a different modification of predatory behaviour – albeit also for peaceful ends. Hounds and terriers used for hunting will complete the entire wolf hunting-sequence right through to consuming their prey, unless they have been trained not to, but this is essentially unreconstructed canid predatory behaviour; the aforementioned hunting elements are present, but subtly reorganized, in the collie. Herding sheep the collie way involves three key elements of predatory behaviour: the ‘eye’ (fixing the gaze, thought to be intimidating), the stalk and the chase. Collie pups start to perform these behaviours at a very early age, and integrate all three into their play. It is then possible for the shepherd to train the young dog to perform each of these separately, on command. The later and most violent parts of the predatory sequence – biting and, eventually, killing – are suppressed if necessary by training, although many collies seem to break off naturally after the chase. Herding ability in border collies is surprisingly heritable (in other words, some collies are born better herders than others), indicating some residual variation even within the breed. Thus selection for this crucial ability, while it must have been and probably still is intense, has not yet reached completion.
A working collie
Among dogs that do not work, many of the skills for which they were originally bred become redundant – or worse – in their companionship role. Sometimes this redundancy does not appear to present any problem; for example, many of the most popular companion breeds – such as spaniels and retrievers – are descended from working animals. But even in these breeds, there has been a tendency for separate ‘working’ and ‘show’ lines to emerge, with most pet animals coming from the latter. This implies that the dogs selected specifically for work may not fit the companion niche as well as they might. The conflict between working traits and the pet owners’ requirements can be even more obvious in herding breeds such as border collies and hunting breeds such as beagles, both of which can require far more exercise and stimulation than the average owner is able to give.
In recent years those who regulate dog breeding have taken on more responsibility to inform prospective owners of breeds’ working origins and the problems these can cause for an owner who is not prepared to adapt to them. For example, the UK Kennel Club describes the border collie thus: ‘He needs a lot of exercise, thrives on company and will participate in any activity. He is dedicated to serving man, but is the type of dog who needs to work to be happy and is not content to sit at home by the hearth all day.’9 Compare this to the official United Kingdom breed standard, which simply says: ‘Temperament: keen, alert, responsive and intelligent. Neither nervous nor aggressive’ – implying to the uninitiated that no collie will become nervous or aggressive, even though some that are denied the active life they crave can become both. (Some references to a breed’s drawbacks are more oblique: ‘There is no better sight than a Beagle pack in full pursuit, their heads down to the scent, their sterns up in rigid order as they concentrate on the chase. This instinct is mimicked in his everyday behaviour in the park: the man with the lead in his hand and no dog in sight owns a Beagle.’10) Nevertheless, prospective owners who take the trouble to investigate the behavioural needs of a breed that takes their fancy should nowadays be able to find fairly accurate information.
However, many of the traits that suit dogs to the companion role – strength of attachment to people, ability to cope with unexpected changes in their environment, trainability and so on – appear to vary as much within breeds as they do between breeds. While not denying that some breeds suit only active lifestyles, whereas others find it easier to adapt to the demands of modern city life, I must emphasize that it is often not so much a dog’s breed as its individual personality that influences how rewarding a pet it will be – and how happy it will be in that role. Breed standards and descriptions can appear to describe a fixed personality (for example, ‘Agile, alert … Should impress as being active, game and hardy … Fearless and gay disposition; assertive but not aggressive’ – from the UK breed standard for the cairn terrier11) but scientific exploration of canine temperament has shown that this cannot be relied upon.
The most comprehensive study of dog behaviour genetics ever conducted was the Bar Harbor project, which started in 1946 and ran to the mid-1960s.12 At the time, psychologists and biologists held diametrically opposed opinions about whether genetics influenced personality: the biologists maintained that many differences in character between individual animals (and people) were influenced by genes, while most psychologists held that personality was the product of an animal’s early experiences. Dog breeds, at that point genetically isolated from one another for half a century or so, were chosen as the ideal starting point for answering such a question.
Five breeds, and crosses between them, were examined for consistent differences in behaviour. The scientists chose small-to-medium-sized breeds with reputations for having contrasting behavioural styles: the American cocker spaniel, the African basenji, the Shetland sheepdog, the wire-haired fox terrier and the beagle. They bred more than 450 puppies, many purebred, but also some that were crosses between two of the chosen breeds. They raised all of them under standard conditions, allowing the effects of any genetic differences to come through. As they grew up, the puppies were given a wide range of behavioural tests. Some of these examined spontaneous behaviour, such as play between puppies in each litter, and the response to being picked up by a person. Others tested how easy it was to train each dog to perform simple obedience tasks, such as walking to heel. Still others tested cognitive ability, such as how quickly it took each dog to learn to get thro
ugh a maze or work out how to pull a bowl of food out from under a wire mesh cover.
A basenji
Surprisingly, when all the results were compiled, breed turned out to be less relevant to personality than had been expected at the outset. Although each breed was found to have some distinctive behavioural characteristics (for instance, the cocker puppies were much less playful than the others), it was only the basenjis that stood out from the rest. Although the descriptions in their breed standards said otherwise, the characters of dogs from the four American breeds overlapped a great deal. For example, although a ‘typical cocker spaniel’ personality could be identified (very dependent upon people, and rather unreactive and low in general intelligence), nine out of the seventeen pure-bred shelties, ten out of the twenty-five beagles and three out of the sixteen terriers (and even two of the sixteen basenjis) also had personalities of this type. Only the basenji, an ancient breed with distinctive ‘village dog’ DNA, were very different in their behaviour. Some of this distinctiveness could be traced to a single dominant gene, which manifested as a tendency among basenji puppies to dislike being handled until they were over five weeks old.
The basenjis aside, many of the behavioural traits that were measured varied almost as much within breeds as between them. The differences between individual dogs were found to be based upon seven different emotional traits (impulsivity, reactivity, emotionality, independence, timidity, calmness and apprehension) and only two traits of ability (general intelligence and the ability to co-operate with people). Whether the dogs performed well or badly in most of the tests had depended not upon differences in their ‘intelligence’ but, rather, on their emotional reactions to the situations they were put in, and their ability to glean clues from the experimenters about what they were supposed to do. Apart from the basenjis, which were much more reactive and inquisitive than the other four, the personalities of all the other breeds overlapped considerably.
Moreover, this study deliberately did not take into account a very important factor influencing a pet dog’s character, its experiences during the first few months of its life. All of the puppies were raised under standard conditions, in order to minimize the effects of such experiences. In the real world in which prospective owners look for pets, such early experience can overwhelm most genetic factors. A cocker spaniel puppy born and raised in an isolated outhouse will behave much like a beagle raised under similar conditions – timid and frightened of anything out of the ordinary – and will be equally prone to developing fearful avoidance or aggression problems later in life.
Some breed organizations place insufficient emphasis on the role of environment in shaping a dog’s behaviour. This is hardly surprising, as they are loath to admit that any of their breeders rears puppies under less than ideal conditions. A dog’s character is the product of a complex interplay between its genetics and the dog’s experiences while growing up. No breed standard for character can protect against the damage done to a puppy by keeping it in an impoverished environment for the first eight weeks of its life.
Perhaps the most important personality trait a potential dog owner will want to know about is whether or not a particular breed is inclined towards aggression. But are there meaningful distinctions between the breeds in this regard? The relative contributions of genetics and the environment in determining whether a dog is likely to bite are still hotly debated: one of the most contentious aspects of personality and breed-specific behaviour is whether aggressiveness is a genetic trait among dogs. It is universally accepted that aggressiveness can be affected by experience, but opinions differ where other contingencies are involved. Many experts now agree that much aggression in dogs overall is motivated by fear, not by anger, and that early experience and learning play a huge role in determining whether an individual dog turns its aggressive feelings into an actual attack. At the same time, however, genetic influences are hard to rule out. In the case of breeds designed for fighting and guarding, they must play a role, although rarely a deterministic one.
Since differences in experience were minimized in the Bar Harbor project, their data should be a good place to look for genetic effects on aggressiveness. All the dogs in the study were tested for aggressive tendencies according to a variety of scenarios, but these did not emerge in the analysis as one of the seven underlying emotional dimensions. Rather, aggressiveness was strongly linked to general reactivity – characterized by a generally rapid heart rate, rapid progress around obstacle courses, and so on. However, none of the five breeds selected for the Bar Harbor project, with the possible exception of the basenji, was especially noted for its aggressiveness, so this study cannot rule out the possibility that genetics may influence aggressiveness in some other breeds, especially those bred for fighting and guarding.
Aggression in such dogs is still an issue of real public concern, despite many measures taken to reduce the risks involved. Except in very carefully defined and tightly regulated circumstances, such as the training of police dogs for public order enforcement, aggressive dogs are unacceptable to most of society. In recent years, attempts to remedy the social problems caused by canine aggression have largely taken the form of breed-specific legislation, much of which has proved difficult to enforce. The laws vary considerably in detail from country to country, but most ban or place severe restrictions on ownership of pit bull terriers and similar breeds. But are pit bulls really different from other dogs, or do they simply have the right ‘look’ for those who wish to use dogs as weapons? The truth probably lies somewhere in between.
Pit bull
Reports of biting incidents are notoriously unreliable,13 so care must be taken in considering whether a dog that has attacked someone was actually a pit bull. Since pit bull types generally lack authenticated pedigrees, pit bulls cannot be called a breed in the same sense that, say, cocker spaniels are. It is thus difficult to identify pit bulls as such; other breeds, most notably the Staffordshire bull terriers, are often mistakenly referred to as ‘pit bulls’. Two other confounding factors may also contribute to the pit bull’s reputation – (a) contagious over-reporting of biting incidents following one well-publicized occurrence, and (b) the deliberate choice of this type of dog by irresponsible owners.
Legislating against a whole breed can be justified only if there are underlying biological reasons why that breed should be aggressive. If, on the other hand, the main cause is irresponsible ownership, manifested as a desire to actually use the dog for fighting (or to merely give the impression of doing so), then outlawing one breed is unlikely to solve anything. Either the breed will be pushed underground, further into irresponsible ownership, or other breeds will take its place.
Pit bulls are certainly descended from dogs intended for fighting. The ancestry of today’s pit bulls can be traced back to bulldogs. Bulldogs were used for bull-baiting in Britain until the sport was made illegal in 1835; after that, they were used for dog-fighting, on both sides of the Atlantic. Such dogs were selected over many generations for specific characteristics: a low level of fight inhibition; rapid escalation of any conflict, often omitting the usual threat communication; and absence of the bite inhibition seen in many guarding breeds, such as German shepherds, which grab and hold but usually do not shake and tear like pit bulls. (The ‘locking jaw’ of the pit bull is, however, a myth.) Breeders reputedly try to select against aggression towards people, for the safety of the owners and their families at least, but it is not clear that such selection is effective: since dogs in general are far more likely to choose their targets of aggression on the basis of experience than through genetically driven preference, they are apt to strike out against anyone they perceive as threatening – including their owner.
Although some pit bulls are unarguably dangerous, many are not. Granted, in 1986, pit bulls were responsible for seven out of eleven fatal dog attacks in the United States, making this breed at least thirty times more likely to have bitten than any other breed. However, none of these seven were
breed-club registered animals, so it was impossible to quantify the effect of their ancestry on their aggression. Furthermore, it is likely that the way their owners had treated them, and especially the way they had trained them, had made the major contribution to their extreme aggression. Despite their fearsome reputation, the vast majority of the 1–2 million pit bulls in the United States at that time had probably never bitten anyone.
The connection between breed/personality and actual biting incidents is imprecise at best, even if one accepts that some breeds are genetically predisposed to be more likely to be aggressive than others, or that there is such a thing as an aggressive personality trait in dogs. Even within so-called aggressive breeds, the dogs that actually attack are extreme outliers. Moreover, the reasons for their extreme behaviour are rarely investigated thoroughly – most are simply destroyed.
In Defence of Dogs Page 32