Although their re-adaptation to the wild offers a compelling glimpse of how dogs might organize themselves in the absence of human intervention, the dingo is a problematic case study. The social behaviour of dingoes has been studied in detail only in captivity, where they maintain a pack structure in which sometimes only one pair breeds. As with wolves, this restriction of breeding could be an artefact of captivity, and of the enforced mixing of unrelated individuals. Moreover, dingoes have experienced several thousand generations of living in the wild, long enough to become undomesticated and to lose the characteristic behaviour of their village-dog forebears. What we know of the dingo is therefore not ideal for understanding how domestic dogs would organize themselves if they were allowed to do so unconstrained by captivity. Better would be studies of dogs that have not reverted to the wild so comprehensively.
Until about ten years ago, the few studies of ferals or ‘village dogs’ that had been published painted a misleading picture of their social organization. Their groupings appeared to be mere aggregations, with little co-ordination of activity. No compelling evidence was found for co-operative behaviour within feral dog ‘packs’. Instead, researchers witnessed dogs fighting over food that could not easily be shared. Similarly, pregnant females would leave their packs to have their young, returning only when those puppies could stand up for themselves, and males played no part in looking after the young.
The reason these early studies of feral dogs revealed such competitive social ‘systems’ was, with a dash of hindsight, obvious. Most of the studies had been conducted in Western countries such as the United States, Italy and Spain, where feral dogs are generally regarded as a nuisance, and are never allowed to settle anywhere for long enough to develop their own social culture. The ‘packs’ that do form are generally composed of unrelated individuals, with none of the mutual assistance from close kin that benefits the typically interrelated wolf pack. Constantly at risk of being shot, trapped or poisoned, and prevented from accessing their food supplies (such as rubbish-dumps), these dogs probably never have the time to establish permanent relationships with others, let alone to develop a culture of co-operative behaviour. To flourish, co-operative behaviour relies on exactly the right circumstances – regular interaction with the same individuals, access to sufficient food and consistent shelter, and a structure based on family ties. Only then will a group be sufficiently coherent to act together against other groups without disadvantaging any of its own members. Because of the environments in which these early studies of feral dogs were conducted, they gave no indication as to where the enormous affability of domestic dogs had originated.
In short, a proper understanding of the sociable behaviour of domestic dogs required studies of feral dogs that were not persecuted, and so could form stable groups without a fear of man. Scientists found what they needed in West Bengal,1 where villagers allow feral dogs to live alongside them, generally tolerating their presence even if the dogs choose to rest just outside a house’s front door. Dogs like those in modern West Bengal appear to be little different from the dogs that lived in the Fertile Crescent of the Near East, where modern civilization is thought to have arisen some 10,000 years ago. Their behaviour may therefore be similar to that of some of the earliest domesticated dogs. In addition to scavenging, these ‘pariah dogs’ are also occasionally given food. But feeding does not constitute ownership; these are independent animals, living a commensal lifestyle like that of a city pigeon.
The independence of West Bengal’s pariah dogs, maintained over many generations, gives them every opportunity to demonstrate the dog’s natural social structure, some aspects of which mirror that of wolves. A given town may support several hundred individuals, but the dogs tend to cluster into smaller family groups numbering five to ten members, just like wolves or indeed other canids. Yet they forage singly, since there is no large local prey available that requires communal hunting, and so it is perhaps slightly misleading to refer to such groups as ‘packs’ in the wolf sense. However, they do share a communal territory, which they will defend against the members of neighbouring groups. So far, so much like wolves.
While some aspects of wild dogs’ social structures are similar to those of wolves, their sexual behaviour and parental behaviour are radically different. Taken as a whole, the reproductive behaviour of these feral groups is quite unlike that of the grey wolf, and much more like that of other species from the dog family with far less structured social lives, such as the coyote. In a wolf pack, only one male and female will breed; they are conventionally referred to as the ‘alpha’ pair. But when each female pariah dog comes into heat, she is courted by many males, mostly from outside her own pack, with up to eight males fighting for her attention. Although she rejects some, several others may be judged worthy, and she is likely to copulate and ‘tie’ with each, sometimes even more than one on the same day. After mating is over, one of these males will often pair up and stay with her until after the pups are born. Some paired-up males will even go so far as to assist with feeding the litter by regurgitating food for them. Whether the male of a pair is usually the father of all, some or even any of the pups is unclear, so his reasons for investing his time in staying with the female can only be guessed at. Perhaps his hope is that if he helps with this litter, she will give him exclusive mating rights next time. Thus every ‘pack’ of pariah dogs annually fragments into pairs, each looking after their offspring separately, until they and their adolescent young rejoin the group. When the next mating season arrives, each adult may form a pair different from that of the previous year. Unlike in wolf packs, there is no apparent consistency in the family structure, nor do young adults help their parents in raising the next year’s litter.
Pariah dogs
Pariah dog packs are, in many ways, quite unlike wild wolf packs – nor do they resemble the artificial wolf packs that are often assembled in zoos. Many groups contain several adult females, and they are usually tolerant of one another, even in the mating season. Unlike captive wolves, none seems to want to monopolize the best of the males, or to prevent others from breeding. Wolf-like ritualized indicators of dominance or submission are apparently never used by either males or females, although the dogs do seem to recognize one another – an ability essential to maintaining a group’s cohesion – and regularly exchange what appear to be subtle greeting signals. However, there does seem to be some sort of hierarchy; the oldest breeding pair is the most aggressive, especially towards any unpaired males. Possibly the male of the pair sees these unpaired males as potential future rivals, while the female may be concerned for the safety of her pups.
A further striking difference is that neighbouring groups of pariah dogs seem to be able to coexist amicably, whereas adjacent wolf packs try to avoid one another whenever possible – and when they do meet, they almost always fight. Although aggression between members of different groups of West Bengal feral dogs does occur from time to time, in many such encounters both dogs defer to each other and then return to their core areas. These dogs do not appear to be motivated by any desire to dominate or displace their neighbours, who must occasionally compete with them for food, even those that they are presumably sure are unrelated to them. In short, the highly competitive nature of unrelated wolves seems to have been completely erased from feral dogs.
The West Bengal studies tell us a great deal about the way that dogs might prefer to organize their lives. They do not seem to be able to adopt the ‘family pack’ structure that is typical of the wolf. Although they do form bonds with family members, these bonds are far looser than among wolves; more important are the bonds between mother (and to some extent father) and their dependent young. As they grow to adulthood, the young share territory with their parents, but do not help them to raise their younger brothers and sisters. While dominance hierarchies are evident, they predict only which dogs get priority of access to food and shelter, not who breeds successfully. So despite the comparatively crowded conditions in w
hich many feral dogs live, they behave like captive wolves no more than like wild wolves. There is not the slightest shred of evidence that they are constantly motivated to assume leadership of the pack within which they live, as the old-fashioned wolf-pack theory would have it.
The affability of dogs who are not related to each other and are not part of the same pack must have been a necessary component of domestication. For instance, domestic dogs live much more closely with one another than wolves do, a characteristic that must be a product of their adaptation to exploiting a new, more centralized food supply. Large predators such as wolves cannot afford to live at high densities, because they would never be able to find enough food. Wolf packs rarely have more than twenty or so members, and they defend very large territories. The dogs that accompanied hunter-gatherers, only recently domesticated from wolves, may have been equally intolerant. However, as soon as humans began to live in large permanent villages, dogs would have needed to evolve ways of coexisting with unrelated animals, the debilitating alternative being constant wariness and frequent fights. Although they exhibit this very characteristic today, and therefore illuminate the domestication process, the situation of the pariah dogs is, needless to say, a long way removed from that of a pet dog living in a typical household. Indeed, pet dogs are usually neutered, and even dogs kept primarily for breeding are not generally allowed to form long-term pair-bonds with the partner of their choice.
Given that pariah dogs do not organize their societies in the same way that wolves do, it seems unlikely that pet dogs would do so either, but the remote possibility does remain that pet dogs might somehow have re-evolved the desire for dominance that characterizes wolves in zoos and is so beloved of some old-school dog trainers. In order to examine this possibility, my students and I worked at a sanctuary for unhomed dogs in Wiltshire.2 Such dogs serve as a useful intermediate between the feral and the pet dog, because they all started their lives in domestic circumstances, before being allowed unrestricted access to other similar dogs. At any one time the sanctuary contains about twenty dogs, usually castrated males whose behaviour towards people is judged to be too unpredictable for them to be homed. At night they sleep in spacious kennels shared by four or five dogs, but during the day they have the run of a large paddock planted with trees and bushes, littered with toys, and equipped with tunnels through which they can run from one part of the paddock to another.
If any group of pet dogs had the opportunity to establish a wolf-type hierarchy, it would be at a sanctuary like this one, where the dogs have eight hours a day of unregulated interaction with their own kind. However, many hours of observation failed to reveal anything of the sort. There was plenty of competitive behaviour to record: dogs would growl or bark at one another, pretend to bite each other’s necks, attempt to mount one another, or chase one another around the paddock. Others would react by crouching, looking away, licking their lips or running off. Yet most of this behaviour took place between a minority of the dogs, whom we ended up calling ‘The Insiders’. Three dogs, ‘The Hermits’, kept out of the way of all the others, and consequently almost never interacted, so it was difficult to know what their ‘status’, if any, might be. Seven others, ‘The Outsiders’, did not actively avoid The Insiders, but always gave way to them. Relationships between The Insiders themselves were inconsistent – it proved impossible for us to construct any kind of hierarchy from our data, let alone any of the types of hierarchy thought to occur among wolves. Over a third of the interactions took place between just four pairs of dogs – Ronnie with Benson (both collie crosses), Jack (a springer) with Eddie (rough collie), Mickey Brown (German shepherd cross) with Branston (collie/spaniel), and Dingo (another shepherd cross) with Tarkus (a Weimaraner). Exactly why these pairs of ‘buddies’ had formed was unclear – none, as far as we could tell, were related – but these attachments were certainly not predicted by any aspect of wolf behaviour. They do, however, emphasize that dogs, unlike wolves, find it easy to establish harmonious relationships with dogs that they are not related to and have not met until both were adult.
Our study of the sanctuary dogs failed to uncover any evidence that dogs have an inclination to form anything like a wolf pack, especially when they are left to their own devices. This reinforces all the other scientific evidence indicating that domestication has stripped dogs of most of the more detailed aspects of wolf social behaviour, leaving behind only a propensity to prefer the company of kin to non-kin – a propensity shared by many animals and certainly not restricted to wolves or even canids. Nevertheless, many experts on dogs and dog training persist in alluding to the wolf as the essential point of reference for understanding pet dogs, despite the fact that the wolf they refer to is not the wild wolf, which values family loyalties above all else, but rather the captive wolf that finds itself in a constant battle with the unrelated wolves with which it is forced to live.
Despite all the evidence indicating that dogs and wolves organize their social lives quite differently, many people still cling to their misguided and outdated comparisons between dogs and wolves. The question therefore has to be asked once again: does the behaviour of the wolf have anything useful to tell us about the behaviour of pet dogs?
The misconception that dogs behave like wolves might not matter if it did not seriously misconstrue the dog’s motivations for establishing social relationships. The most pervasive – and pernicious – idea informing modern dog-training techniques is that the dog is driven to set up a dominance hierarchy wherever it finds itself. This idea has led to massive misconceptions about their social relationships, both those between dogs within a household, and those between dogs and their owners.
Every dog, conventional wisdom holds, feels an overwhelming need to dominate and control all its social partners. Indeed, the word ‘dominance’ is used widely in descriptions of dog behaviour. Dogs that attack people they know well are still universally referred to as suffering from ‘dominance aggression’. The term is sometimes even used – incorrectly – to describe a dog’s personality. Consider this quote from the American Dog Trainers Network website: ‘A dominant dog knows what he wants, and sets out to get it, any way he can. He’s got charm, lots of it. When that doesn’t work, he’s got persistence with a capital “P”. And when all else fails him, he’s got attitude.’3 Actually, this is just a description of an unruly, untrained, yet somehow charming, dog. It says nothing about what that dog’s relationships with other dogs are like – nothing about its relationships, ‘dominant’ or otherwise. Other inaccurate or misleading uses of the term ‘dominance’ abound in dog training. For example, celebrity trainer Cesar Millan has referred to dogs trying to ‘dominate’ cats, and even a dog chasing the light from a laser pointer has been described as trying to ‘dominate’ it,4 behaviour that a biologist would immediately classify as predatory, not social.
Used properly, ‘dominance’ means something quite different from these meanings assigned to it by dog trainers. The term simply describes a relationship between two individuals at a particular moment in time; it makes no predictions about how that situation arose, about how long it is likely to last, or about the personalities of those individuals. Indeed, a biologist would be likely to point out that if you put the dominant one of the two in another social situation, it might well end up not being ‘dominant’. Moreover the term is just a description – it says nothing about whether the two animals involved are in any way themselves aware that theirs is a ‘dominance’ relationship.
Calling dogs ‘dominant’ also suggests that their relationships can be fitted into hierarchies. The concept of ‘hierarchy’ is the corollary of ‘dominance’, whenever a group consists of more than two individuals. If each pair can be observed to have a dominance relationship, and if those relationships can be arranged in a linear fashion, such that each individual is subordinate to all those above it, a hierarchy can be determined. Sometimes no such hierarchy can be constructed; this is the case, for example, when individuals that
should be ‘low down’ in the hierarchy according to most of their relationships nonetheless ‘dominate’ an individual that ‘dominates’ many others and should therefore be ‘high up’ in the hierarchy. Furthermore, just as the dominance relationships between all the pairs that make up the hierarchy may not be appreciated as such by the individuals involved, so too the hierarchy may not be visible to those within it, even if one is evident to an outside observer.
There is little evidence that hierarchy is a particular fixation of dogs, either in their relationships with other dogs or in those with their owners. Let us first examine what hierarchy is. Let us assume that dominance relationships are based on how much of a single attribute – say, fighting ability – a dog possesses: the dog with the most of that ‘quality’ should be dominant over all the other dogs in its household, and might be labelled the ‘alpha’. The dog with the second largest quantity of fighting ability should be dominant over all the others, except the alpha, and is conventionally labelled the ‘beta’. The dog with the least fighting ability – the one subordinate to all the other dogs – is labelled the ‘omega’. In a four-dog household, the hierarchy might look like this:
Linear hierarchy
Relationships between real animals are usually not as simple as this neat hierarchical arrangement. Some relationships may have been resolved by fighting; others, through superior problem-solving ability. Still others may be historical in that the ‘subordinate’ dog was brought into the household as a puppy and has never challenged any of the older adults. When such complexities apply, it may be possible to distinguish consistent relationships between pairs of dogs, but impossible to arrange them in a hierarchy, like this:
Circular ‘hierarchy’
Traditional thinking presumes that the dog’s hierarchy stems from the wolf’s, and more specifically, from that of the captive wolf – but we now know that this conception is fundamentally misguided. The typical structure of a captive wolf ‘pack’ was thought to be made up of two hierarchies, one for each sex, although in larger packs these are not entirely linear, since dominance relationships between the juveniles and between the cubs are often indistinct. However, these ‘hierarchies’ are now thought to be an artefact of captivity, whereas the structure of a natural pack depends largely on family allegiances in which no aggression-based hierarchy is apparent at all.
In Defence of Dogs Page 10