The Culture Clash
Click on the cover above to go to this book at Amazon.co.uk
Jean Donaldson's aim in 'Culture Clash' was to market the ideas of B.F. Skinner, a key behaviorist pioneer who developed the notion of operant conditioning. Skinner was interested in how animals reacted to rewards and punishments, rather than what was inside their heads, what they felt, or what their motives were. Donaldson tells us on p10 that the behaviourist model is 'inescapably verifiable'. Her book became a best-seller, and is still on the list of recommended reading in some modern training books. This is quite a feat, given that only the last chapter is an explicit 'how to' guide to training, and it's difficult to find training points in other chapters, amid the diatribes!
Since Culture Clash was written, there’s been an explosion of interest in what goes on inside dogs’ heads, partly from ethologists, specialists in animal behavior. Initially, they were interested in wild animals in their natural environment. Then a Hungarian, Vilmos Csanyi, directed a team to study canine ethology, arguing that, because dogs have always been with humans, domesticated dogs are in their natural environment. Research into canine ethology has blossomed worldwide. Ethologists note that dogs are successful because of their close relationship with us. They're more versatile than wolves, and have a longer socialization period, so can learn more. Most dogs are smart at assessing humans, they like doing things with us, so need to understand us. They’re better at understanding our body language than we are at understanding theirs. Dogs are social animals, so have social rules, or 'morality', though it's not quite the same as human morality. Meanwhile, psychologists became interested in canine cognition, moving away from behaviourism, and sometimes collaborating with ethologists. A strength of this newer research is that it looks at dogs, rather than applying research on rats and pigeons to dogs. It also underlines the deficiencies of 'Culture Clash' as a source of information on what makes dogs tick.
Trainers have to become skilled at reading canine body language to guess a dog’s feelings and intentions. If they don’t develop this skill, they’re more likely to be bitten! In some ways, academic scientists are starting to understand issues that people with hands-on experience have long understood, if intuitively.
So what was 'Culture Clash' about? Chapter 1, 'Getting the Dog's Perspective (Dog Intelligence and Morality)' begins with a tirade against simplistic ideas of dogs as wolves, and the portrayal of dogs as intelligent and moral in the same way that humans are. This chapter is sadly uninformative. Over-simplistic views are wrong, whether it's seeing dogs as 'intelligent', with human intelligence as a yardstick, or dismissing them as 'lemon brains', as Donaldson does here. Likewise, her portrayal of dogs as amoral ignores their sense of social norms.
Chapter 2, 'Hard-Wiring: What the Dog Comes With (Predatory and Social Behavior)', makes some useful points, like the benefits of letting dogs play to learn social skills, and of tug games with rules, as a reward in themselves, and to enhance communication and control. Donaldson does, however, tend to downplay inbuilt differences between dogs. Some dogs learn very quickly how and when to solicit play, and how to regulate their strength. Others need more training, supervision, and careful selection of companions. Owners need to learn that not all dogs welcome their dog's attentions, as well as when to allow play, and when play turns into bullying. It's normal to object to unmannered, rough 'play', whether you're a dog or a human.
Donaldson also plays down inbuilt differences when she discusses retrieving. Given similar training regimes, some dogs can become obsessive retrievers, while others have less interest. It would be helpful to mention 'off switches' or cues, and ways to channel that desire to higher level work. Donaldson also presents the desire to 'bite/hold/shake/kill' as inbuilt, but again it's much stronger in some dogs than others, and such dogs also need an 'off switch'.
Donaldson sees 'separation anxiety' as inbuilt. She prescribes desensitization, and getting another dog. True, these remedies may work, another dog may help a rescue greyhound, used to being with other dogs. However, being very attached to a human is common in breeds designed to work closely with humans. Getting another dog won't help if a dog wants human company. It can also create problems if the dogs aren't compatible. Some dogs calm one another, others get each other overexcited, or hate each other; bad news if they're alone together all day. Furthermore, dogs may bark and trash the house for reasons other than separation. Rehomed dogs may panic simply because they're in an unfamiliar place with a new routine, and they don't know what's happening. Dogs left alone for long periods may trash the house out of boredom.
Donaldson downplays the importance of a dog's age (p35), yet this is critical for socializing dogs to humans, other dogs, and small domestic animals, like chickens. Orphan pups, for example, need early contact with their own species, to learn social skills. Again, much of the chapter is taken up with diatribes, which leave little space for information.
Chapter 3 is called 'Socialization and Conflict Resolution (Biters and Fighters)'. Donaldson states that spooked dogs have to choose between biting or fleeing, and stresses socialization to ensure that dogs aren't spooked. In fact dogs have a wider repertoire than bite or flee in response to threats. They may freeze, try appeasement gestures, or seek protection from their owner. Socialization is important, but so too is building trust so dogs can learn to use this wider repertoire. Donaldson's advises allowing pups to bite their owners, though not young children. She claims that little pups can't learn to hold back on bites unless you use 'harsh punishments' (p69). Yet I've taught pups very quickly not to bite, by blowing a gentle raspberry into their faces when they were lying in my lap and tried nibbling me. They stopped, started licking me, and never used their teeth on me again.
Donaldson stresses that socializing dogs with strange humans is very important, which is true, though her 'socialization hit list' on p63 portrays socialization as getting strangers to give your dog cookies – not always sensible. Dogs do need to learn that the default approach is to focus on you, not the stranger, and to go up for a treat only if you give permission. Donaldson also downplays temperament, but it's important, especially if a dog is expected to handle an environment with a lot of people coming and going, including children. Choosing a 'sociable breed' is a start, but some dogs from popular 'family' breeds have dodgy temperaments, so it's wise to check a pup's line.
Chapter Four is called 'It's All Chew Toys to Them (Behavior Problems and Solutions’ - an odd title, in that chewing is a less serious problem than biting). Certainly, Donaldson is accurate in saying that dogs don't see objects as we do. However she underestimates their intelligence when she argues that they have no perception of the value of items to owners, but just see them all as chew toys. Dogs are good at checking us for information, and can deliberately select 'forbidden' items in order to gain owners' attention! Training has to go beyond just teaching what it’s OK or not OK to chew.
Chapter 5, 'Lemon Brains But We Still Love Them (How Dogs Learn)', presents dogs as unintelligent, only able to learn through rewards and punishments. Donaldson exhorts owners to learn the 'basic rules', that rewards and punishments can be given or taken away. She makes some useful points, for example that dogs can be allowed rewards that fit their natural desires, and that feedback, and its timing are important. Her description of commands as building bricks in training is also helpful. One of Donaldson's examples of rewards is a gambling machine. She tells us that we can't help but get hooked on these machines, because they're designed using 'laws governing reinforcement' (p144). That's a bad example. Go into a bar and watch clients and slot machines. Some people are addicts, others are immune. Humans, like dogs, have built-in differences. There's no 'law' that tells us what's rewarding for a particular human or dog, nor is it clear whether the slot machine 'reward' is in the payout, or in the excitement of anticipation. Unravel operant conditioning, and it's much less clear-cut than it first appears.
Donaldson claims that punishment doesn't kill behaviour, what kills behavior is removing the reward. Punishment just temporarily stops the dog (p159). This is simply not true, punishment can be very effective in permanently stopping certain behaviors. If animals didn't learn from punishment, we'd be at risk of extinction as a species. She in fact later states that positive punishment (adding something to stop behaviour) works under certain conditions, and she argues that you should use a 'large punishment' the first time a dog does something you don't want (p162). Yet very mild aversives are often more effective. As with medicine, a stronger dose isn't always more effective, and can be counter-productive. Furthermore, 'punishments' can be qualitatively different. Hosing down fighting dogs is more likely to interrupt a fight than hitting them, which may simply encourage them to fight harder, and bite the human.
Donaldson also fails to spell out that withholding expected rewards is a form of punishment in operant conditioning terms, 'negative punishment', or taking away something to stop a behaviour. She believes in withholding rewards, so she sees punishment as effective, though she’s fostered a myth that negative punishment isn't really punishment. Some of her advice on providing rewards is also unhelpful, like rewarding a dog for the absence of unwanted behavior (p163). How does the dog know what the reward is for? Why not reward a dog for obeying a command in the presence of a distracting stimulus?
There are good reasons for trying to provide rewards rather than inflicting punishments. Rewards tend to boost trust, punishments can undermine trust, and a dog's response to punishment isn't always predictable. A dog hit by a car may avoid cars, or chase them because they're dangerous. Attempts at either positive or negative punishment may be counter-productive. Focusing on providing rewards makes for a better relationship and is more fun for dog and owner. However, trying to avoid punishment altogether isn't sensible. If owners develop two-way communication and trust, they're more likely to have an intuitive understanding of what's effective with a particular dog, and the dog's forgiveness if they get it wrong.
Behaviourism doesn't of course give a complete picture of how dogs learn, and this is clear from gaps in Chapter 5. First, as in the chapter on hard wiring, there's no mention of sensitive periods, when dogs need to learn key skills, or teaching them becomes very difficult. Likewise, it's more efficient to start teaching high-level skills when dogs are young. Structured obedience and skills training programs are also more successful if they take into account the dog's level of maturity. Furthermore, there's no mention of social learning, that dogs can directly teach each other both good and bad behaviour, and can learn by imitation, from observing each other and their owners.
Donaldson also claims that dogs only have a desire to please if we teach them they get to do what they want if they obey us. This misses the point that what dogs often want is our company. They're far more likely to take note of 'I'd rather you didn't do that' than are wolves. True, adolescent dogs tend to be more independent and often need work on impulse control. However, as dogs mature and they and their owners get to know one another, some dogs do develop a mature 'desire to please'. Relationships, and learning how to communicate, take time and effort to develop.
This chapter often saddens people with working dogs. The list of what owners need to learn, on p168, doesn't mention encouraging dogs to communicate. Owners need to listen, if only to learn when a bark means 'a dog has passed by', and when it means 'let me out immediately, I'm going to expel disgusting substances from both ends'. People who work their dogs develop two-way communication and know that their dogs are, in many ways, smarter than humans. A shepherd and his sheepdog, or a blind person and their guide dog, see the relationship as a partnership. Yes, the human is the senior partner, taking the executive decisions, but dogs can understand some things we can't. Sheepdogs sometimes predict what sheep are about to do better than the shepherd. At times, the dog takes over leadership. These dogs certainly aren't 'lemon brains'. Furthermore, dogs are carefully selected for high-level tasks, there has to be an inbuilt desire to do the job. When a dog really wants to do a job, a 'correction' may be welcome because it conveys information that the dog needs.
Skinner focused on observable behaviour, not what an animal felt, so 'punishment' became just something that stopped a behavior, or made it less likely to happen. Skinner's pigeons and rats avoided electric shocks. Most people would guess they were upset by them, though guesses and assumptions about the animal's feelings are unscientific in behaviourist terms. Today we can measure hormonal and other changes generally associated with distress, so make more accurate guesses about how dogs feel when humans do something that stops a behaviour. We also know that interactions between dogs and humans are complex, and that dogs look to us for information. When a correction conveys information that the dog wants, it can end a behaviour without being an aversive. Furthermore, it can be distressing for a dog to ask for information and not get an answer.
The last and sixth chapter deals with obedience training, and again has some good training points, especially for people interested in clicker training. Obedience training is basically teaching dogs self-control, especially sitting, staying, and recall, basic commands that are important if you want to be able to trust your dog in most situations. Self-control is important, but if owners restrict themselves to control, and fail to develop two-way communication, they miss out on much of the joy of dog ownership.
A key gap in this book is a serious discussion of the why of training. Donaldson holds up marine mammal trainers as shining examples of practitioners of behaviorism (p129). But why deprive a wild animal of its freedom to teach it tricks to amuse us? Why shrink from using mild aversives if they help keep a dog alive, and mean we can give the dog more freedom? A sad effect of an obsession with 'punishment' is that we can lose sight of these wider issues, which include making sensible choices. We can connect to dogs in ways that we can connect to no other species. They don't kid themselves in the complicated ways that we do, so can teach us about ourselves. They tell us a lot we can't perceive because of they can smell and hear better. But to enjoy canine company, we need to be sensible about whether we really have the time for a dog. A dog’s temperament also affects whether we enjoy their company, not so much 'good' or 'bad' temperament as a dog with the personality and energy levels that suit ours. People who don't make sensible choices are more likely to abandon their dogs. Donaldson ends her book telling owners to spay and neuter their dogs. Yet in some countries, especially in Scandinavia, dogs are rarely neutered, and unwanted dogs are less of a problem. Perhaps Scandinavians make more sensible choices.
Donaldson succeeded in marketing a particular 'brand' of dog training. She appeals to human emotions with bang-on-the-table arguments, rather than careful reasoning and meticulous observation. She has helped foster a 'two camps' model of dog training. Humans have a built-in tendency to divide the world into 'us and them', yet both dogs and humans can learn to be flexible about who's an insider, and who's outside our group. Talk to dog people individually, and you can listen to quieter people who are often shouted down. Your understanding can be enriched by listening to people with different experiences. A simple 'two-camps' view of training is an impoverished view which taps into our biology, but doesn't really help us understand dogs.
'Culture Clash' is a landmark book which can help you understand the history of professional dog training. It isn't a scientific text, nor does it tell you much about what makes dogs tick. Donaldson 'sold' behaviourism, without explaining it clearly, or spelling out its limitations. Zulch and Mills’ ‘Life Skills for Puppies’ is a good ‘how to’ book for those critical early months that Donaldson neglects. Patricia McConnell’s ‘The Other End of the Leash’ is more helpful for understanding a relationship between animals of different species, humans and dogs. Burch and Bailey's ‘How Dogs Learn’ gives a clear account of operant conditioning. Steven Lindsay's ‘Applied Dog Behaviour and Training’ offers a thorough approach to training from someone with dog sense and an understanding of scientific research. For an account of recent scientific research on dog behaviour, try ‘Dog Behaviour, Evolution and Cognition’ by Adam Miklosi, who now heads the team that Csanyi set up. Then check what these authors say with what real dogs tell you.
Review by Alison Lever
Thank you to Wendy Hanson, Tiffani Howell, Heather Houlahan, Donald McCaig and especially Janeen McMurtie for helpful discussions about behaviourism in dog training
Further Reading
Burch, M.R and Bailey, J.S. (1999) How Dogs Learn. Howell Book House
Lindsay, S. (2005) Handbook of Applied Dog Behavior and Training, Volume 3: Procedures and protocols. Blackwell Publishing.
McConnell, P. (2002) The Other End of the Leash: Why we do what we do around dogs’. Random House.
Miklosi, A (2014) Dog Behaviour, Evolution and Cognition. Oxford University Press
Zulch, H. and D Mills (2012) Life Skills for Puppies. Hubble and Hattie
