If Dogs Could Talk: Exploring the Canine Mind


 Click on the cover above to go to this book at


In the scientific community, there are two types of knowledge concerning dog research: things we have proven through rigorous experiments, and stuff we "just know". 

As a fairly recent addition to the field of dog research, I have witnessed firsthand the stark contrast between what is written up in a scholarly journal article and what is spoken behind closed doors. Vilmos Csanyi’s If Dogs Could Talk elegantly combines the two kinds of knowledge that we have about these beloved animals.

In this book, Csanyi takes great care to inform the reader about the scientifically-proven information we have gained about dogs, and is diligent in drawing a line between scientific studies and anecdotes, while also noting that, taken together, large numbers of anecdotal observations can sometimes eventually become accepted in the body of scientific knowledge. After all, his own field of expertise, ethology, is essentially the observations of animals in their natural habitat until conclusions can be formed about various behaviours of the animal.

He successfully demonstrates the rather obvious fact that the average dog owner understands far more about dogs than is currently permitted by science. Because science won’t allow attribution of human emotions to animals, dogs cannot "love" their owners, but instead "form attachment bonds". The irony is that many dog scientists are also dog lovers and know full well the love that can exist between a dog and its owner. Csanyi bravely stepped into the public eye and bluntly illustrated this disconnect.

The best thing about this book is the mixture of hard science and dog knowledge that is yet to be scientifically proven (I say "yet to be" because some of what we "just know" about dogs will be proven scientifically in a matter of time, as the field goes from strength to strength). Coming in at a close, photo finish-style, second place are the adventures of Flip and Jerry. By the end of the book, I was dying to meet these two characters! Their behaviours and personalities – I can say that because the scientific community has started to accept that dogs do have personalities – shone through in every story he relayed, and it made me wonder what fantastic and weird things my dog has done that I have forgotten about simply because I have not kept records of every time she did something interesting or funny. Csanyi’s meticulous record-keeping made the book light-hearted when it could have become cumbersome, and I’d wager that it made his own efforts at writing the book a lot easier than if he had tried to remember everything later on.

My only criticism of the book is the last chapter, and even then only a few sentences. In his effort to explain the best way to be a dog owner (not scientific!), most of his suggestions are plain common sense. However, the breed-specific comments made me raise an eyebrow, as did the references to dog intelligence tests. Breed-specific legislation is becoming more and more common and lacks a scientific basis. In fact, many breed-specific studies have shown that there is often more temperament variability among individuals within a breed than between breeds. Perhaps he would have been better served to illustrate the public’s perception of these breeds, and highlight the importance of further studies to establish if there is a link between dog bites/attacks and certain breeds. However, those were about two paragraphs out of 300 pages of fascinating and thought-provoking information about dogs.

To conclude, Csanyi was courageous enough to say aloud what many dog researchers believe – the intersection between what we’ve proven and what we know. It is highly readable and, most importantly, a lot of fun.


If Dogs Could Talk: Exploring the Canine Mind

Vilmos Csanyi 

Review by Alison Lever

Vilmos Csanyi is chair of the Department of Ethology at Eotvos Lorand University, Budapest, Hungary, where a team, which also includes Adam Miklosi, has been studying canine ethology, evolutionary systems, and behavioural genetics for over a decade. If that puts you off as very academic, relax, this book is chatty, enthusiastic, and full of anecdotes about Csanyi's own dogs. If it intrigues you, and you want a meatier, more academic account, try Miklosi's book.

The first part of the book deals with the origins of dogs, how they developed from wolves, and differences between dogs and wolves, differences which are explored further in the second part, which focuses on the characteristics of dogs. Csanyi then goes on to explore ideas about communication between humans and dogs through excerpts from his diary about his own dogs, Flip and Jerry. The fourth part is perhaps the most challenging for non-biologists; Csanyi traces the history of the study of animal behaviour, mentioning Darwin, Pavlov, behaviourism, and what he calls the 'new anthropomorphism'. Lastly, Csanyi looks at dog ownership and breeding in a single chapter.

Dogs are not wolves

The first part of the book is especially useful for people interested in similarities and differences between dogs and wolves. The team has raised tame wolves, as well as researching genetic differences, and they have found that wolves do not bond with humans as dogs can, if they have been exposed to humans between 4-12 weeks. Wolves have a shorter socialisation period than dogs. The longer socialisation period of dogs gives them more chance to bond with humans. Wolf pups are more attracted to other wolves than to humans, while dog pups are more interested in humans than other dogs, apart from their mothers. Wolves are not interested in human speech, whereas dogs are. Wolves don't stop what they are doing on command, whereas dogs can be trained to do so - dogs are better equipped to obey human rules. Wolves don't 'switch off' when left alone, but look for something to trash, whereas dogs tend to snooze. Wolves are also more likely to fight each other than are dogs - it is possible to keep large numbers of dogs together without fights, but not wolves. All this is a useful reminder of the limitations of models that see dogs as 'designer wolves'. Furthermore, Csanyi suggests that modern wolves have become more timid as a result of being hunted by humans, which makes it even more problematic to see dogs as 'designer wolves'. I did miss a longer discussion of one further difference, that dogs are less xenophobic than wolves, ie less aggressive with strange dogs and humans, as well as those they live with. This is just as well, or we would never be able to walk dogs.

Dogs share some characteristics with humans

Part 2 focuses on dogs, and ways in which they have become more like humans in their behaviour, a change made possible through bonding. Csanyi notes, for example, that dogs can be self-regulating, as humans can. You can teach a pup to wait until permission is given before it eats, just as humans can learn to hold back, so longer as owners are consistent in applying rules.

There is general agreement that dogs feel certain emotions, such as fear and joy. Csanyi suggests that they may also feel empathy, become anxious for their owners, and may even feel jealousy, but certainly do not share the emotion of disgust with humans. The way dogs respond to human emotions of course varies. Some dogs are more emotionally robust than others, and some humans 'read' their dogs and communicate with them better than others. Dogs may, as Csanyi suggests, feel empathy, but not all dogs do, just as not all humans do. My mother's younger border collie was upset by heated discussions between humans, and used to hide behind a chair if humans appeared to be too antagonistic. Her older sister once went up to a weeping friend and nipped her. Maybe she was saying 'pull yourself together woman', maybe she just wanted a peaceful, orderly life, and was 'herding' this woman into peaceful behaviour, who knows. My collie-keeshond cross was unperturbed by heated discussions. And one day, when I was weeping after I had heard of the death of a friend, he came up and laid his head on my lap.. Was this empathy? who knows. But it had the desired effect, I took him out for a walk and felt better.

Dogs can obey rules, are are sensitive to human rank order. Csanyi notes. He gives an example of colleague's dogs obeying him because he was the boss, but not obeying other colleagues. Csanyi does not go into rank order in multi-dog households, which might affect how what he calls 'jealousy' is expressed. One of his examples is of dogs from a local pound, used for experiments. The dogs enjoyed being singled out to work with experimenters, but were picked on by other dogs if they were especially favoured. He also says that if one dog is favoured too much, another will grumble. I have found that sometimes the favoured dog causes trouble, if it is a low-ranking 'pushy' dog, emboldened by owner attention to take on the highest ranking dog. This happened with Toby, an ex-outdoor hunting-dog mutt, around three years old, who was dumped by his owner and lived a few weeks as a stray before coming to live with me. He was not used to being petted, and was so emboldened by a scratch behind his ears that he tried to take on Conor, the top dog, and oust Conor from my lap.

Again groups of dogs vary a lot in how they interact, as do humans. I surveyed multi-dog households, asking owners whether or not they perceived a rank order, and if so, how it was expressed. There was a great deal of variation, both in whether a rank order existed, and in how owners used it to control the dogs. Likewise, some dogs and some humans are more inclined to obey rules than others. Csanyi quotes a study comparing basenjis to shelties, to see whether pups from these breeds would refrain from eating from a dish of food after they had been told not to. The pups obeyed while the experimenter was in the room. As soon as he left, the basenjis tucked in, while the shelties continued to respect the prohibition.

All Csanyi's generalisations, then, have to be filtered through the caveat that dogs vary in their behaviour. Genetic factors influence the extent to which dogs obey rules, as does training. Dogs in general are happy to avoid conflict by obeying rules, but some dogs need more reminding than others.  

Csanyi stresses that dogs like to undertake cooperative activities with people. They play many important roles, such as service dogs, tracker dogs and sheep dogs. He was puzzled that wolves appeared better able than dogs to solve some problems, while dogs can carry out tasks needing a high level of intelligence. The reason why emerged not from a comparison between dogs and wolves, but between house dogs and outdoor dogs. The dogs were set a task which they first had to solve independently, and then with encouragement from their owners. The outdoor dogs initially performed better, but the house dogs did just as well once they had received permission from their owners.

Some people might see waiting for permission as 'learned helplessness'.. It can also be called 'self-discipline' and accepting the rules. Many years back my front door blew open, and the three dogs indoors could have taken themselves for a walk. I was very grateful that they had understood the rules and hadn't. They were just snoozing unconcernedly. Obeying the rules had kept them safe. Independent dogs become street-wise, but can also get into trouble. Free-ranging dogs in the village where I live in Spain are often killed by cars. Independent dogs are also less able to enjoy co-operative activities with humans, until they have learnt more about obeying rules and communicating.

Co-operating in complex activities demands a lot from dogs. They have to understand the overall task (eg helping a blind owner to get from A to B, or bringing sheep down from a hill), break down the task, exchange information with their handlers, and alternate dominance. While in general terms the blind owner or shepherd is in charge, deciding what the task should be, there are times when the dog knows more than the human, and, even at the risk of being chastised for disobedience, will take over. A dog may, for example, steer a blind owner away from a pothole, because the overall task, getting the owner safely from A to B, is more important than the owner's insistence that they stick to a very specific route. This raise interesting questions about how much dogs understand of what goes on in human minds. For a dog to understand that 'I can perceive this and my owner cannot', and furthermore, to have the ability to tell the owner he is wrong, while accepting that the human is the overall leader, is no mean achievement. Not surprisingly, it takes a long time for handler-dog teams to build up high levels of communication and trust. Csanyi mentions that in the best performing teams, handlers often scolded dogs for apparent disobedience, and then apologised when they realised the dog was right.

The apparently seamless alternation of leadership between working dogs and handlers, their apparently effortless ballet, involves owners understanding canine signals such as nose nudging, barking, and glances to point at an object. Part 3, The Diary of Flip and Jerry, contains many examples of Csanyi's dogs giving information and also asking questions. He muses that perhaps his dogs ask so many questions because they are ill-mannered. But, in the words of George Hobson, gundog trainer, 'a good leader knows how to listen'. A well-mannered dog can ask questions politely. Flip and Jerry can also role play, for example pretending to be fiercely defending an object. And they can imitate human behaviour, if trained to obey a 'do as I do' command.

Is anthropomorphism a 'bad thing'?

Part 4 is very interesting for people trying to make sense of fads in dog training which claim to be 'scientific'. In the US, in particular, dog training and books on training have tended to be dominated by behaviourism, a branch of psychology that has focused on observations made in laboratories rather than in nature, excluding genetic diversity from its area of study. Also excluded are canine insights, consciousness, thinking and emotions, because these are difficult to study. Behaviourists argued that attributing emotions or desires to animals was anthropomorphic. Younger behaviourists went from neglecting higher order phenomena to claiming that they did not exist. At this point, Csanyi argues, behaviourists lost touch with reality. For many dog owners, Jean Donaldson's 'Culture Clash', just did not 'fit' with what they had observed in their dogs. It was not just that there was a lot that behaviourism didn't explain, sometimes writers got carried away, and were just plain wrong.

As if this were not enough, some trainers and writers from the US claim that positive reinforcement is the only way to train dogs. Confusingly, they call their methods 'pure positive' though strictly speaking, positive punishment would also fit into a 'pure positive' description'. There is no doubt that corrections that are too harsh are a 'Bad Thing', but ... trying to raise and train dogs without ever doing things they don't like, is just plain daft, and not at all 'scientific. One proponent of 'pure positive' methods privately justified his writings on the grounds that his Amazon ratings were high - not on the grounds that they were scientifically respectable.. The term 'pure positive' has everything to do with marketing, and nothing to do with science. What can one counter it with, that one believes in 'impure negative' training. However, Cesar Millan has recently erupted on the scene preaching a very different message, and has shot up Amazon charts. He has countered the marketing message of extreme and dumbed-down behaviourism with his 'natural' view of dogs. 'Natural', like 'pure' is a word many people would not argue with.

Curiously, Millan claims to base his ideas on ethology, on 'nature', and inbuilt tendencies of dogs to do certain things. As he and Csanyi both see ethology as important, there should be some 'fit' between their ideas. Is there? To some extent yes, both see it as important that owners take dogs for walks, and both stress that owners should not be afraid to correct dogs, and should not allow them to get the upper hand. But Millan focuses much more on 'brattish dogs' using ideas from now very outdated studies of wolves. He tends to emphasise dominance, and plays down training, consistent teaching and enforcement of house rules. He also tells a very partial story which misses out on the potential for communication between dogs and humans, and the wide range of activities that dogs are capable of. As for treadmills, well, as people used to say in my home town, 'what is the point of getting a dog if you aren't going to take it for walks'.

Both Millan and behaviourists oppose what they see as anthropomorphism, except that Millan tends to call it treating dogs like humans. And, while extreme anthropomorphism is obviously a Bad Thing, very mechanistic views of dogs that see dogs as responding to stimuli, and communication as basically one way, from us to them, are just as extreme. New anthropomorphists argue that it is possible that humans and animals share mental characteristics, but theories should be verified with experiments and observations before they are taken seriously.

So science can go up blind alleys, and then 'dog gurus' can squeeze even further up those alleys, by a crude popularisation of what was already a little dodgy to start with. Csanyi is disarmingly frank about ways in which academic scientists, who are human like the rest of us, can be blinded by fads. Anyone who has trawled through academic journals to look for articles on dogs can find some very bad science. What then is the point of science, if scientists can get it wrong? To quote Csanyi 'the correctness or truth of a theory is proved not by its beauty or simplicity, but by its practical value' (p.193). Scientific methods, starting out with a sensible hypothesis, making careful observations, rejecting theories that turn out to be unhelpful, and developing those which can be used to make predictions, are useful in any context, both inside and outside the academic world. What helps make Csanyi a particularly good researcher is that he used to herd sheep with a puli as a child. He already knew that there was something going on on a dog's head that was worth researching.

Oh dear, Vilmos ...

Csanyi, then, is a very interesting researcher, but .... where this book is particularly weak in in part 5, 'How to Be a Dog Owner'. Every 'dog person' I know who has read the book comes to this chapter and says 'Oh Dear', or words to this effect. Csanyi, we love ya but ....

On the plus side, I like his empasis on commitment, and the recommendation of an hour or two's walk a day seems sensible, for a UK owner. He makes common sense comments on corrections, tailoring them to the dog, for example. And then he tackles breeds and breeding. A lot of space is devoted to Stanley's Coren's dubious views on breeds. Csanyi mentions that intelligence is difficult to measure, since motivation and training can affect performance. So if Coren's views are of limited value, why reproduce them?

Csanyi himself likes mutts, though he recognises that some people like pedigree dogs for their predictability. Perhaps he could stress that this is relative. Breed does not determine genetic inheritance, nor does genetic inheritance determine behaviour. There can be a great deal of variety within breeds, and a particular pup's line may be especially important for owners wanting predictability, while training is obviously important. Csanyi is especially unhelpful when he argues that dogs from 'dangerous breeds' should not be chosen as family pets. Some breeds do need especially responsible owners, if only because of the publicity that results when things go wrong.. As a scientist, perhaps, more useful advice would be to research breeds thoroughly, by meeting real dogs from that breed, and if possible from a pup's line, before making a choice. And, as far as training goes, to ask people with well-trained dogs how they achieved their results. Some of the best trained dogs I have met are owned by people who have never opened a dog training book, let alone been to a training class.

Part 5, then, could do with a rewrite, especially for a UK and US readership. Whatever the position in Hungary, UK and US owners are concerned that careless remarks by prestigious scientists could fuel calls for yet more breeds to be banned - a point that Csanyi's colleague, Adam Miklosi, is well aware of. Teaching house rules is especially important in the UK and US, as well as teaching dogs to behave well with strangers. This is why a discussion of xenophobia would have been useful in part 1. Dogs can learn to treat people they meet regularly as part of their 'pack', and to accept anyone their owner is friends with as a 'pack member', but it has to go further than this. Any tendency towards xenophobia, aggression towards strange humans or dogs, has to be curbed, because the consequences of not doing so are serious. They are serious not only for individual owners and dogs, but affect the public perception of dogs, so owners have a social responsibility to teach dogs to behave well with strange humans and dogs. Curbing xenophobia does not have to be difficult, usually it is simply a question of supervising early encounters with strangers. A stress on the commitment involved in taking on a pup is also important in those countries such as the UK, and especially the US, where many dogs end up in shelters. What would have been helpful for all dog owners is a guide to activities that owners can do with their dogs, based on Csanyi's own research. A mental work-out can help 'settle' a dog as much as a long walk, and teaching dogs complex activities makes life more fun for both owners and dogs.

On balance, this is well worth reading

So, this book does have flaws, but it's still a very readable and interesting book, which perhaps merits a new edition with those flaws corrected. The team that Csanyi directed towards canine ethology is now headed by Adam Miklosi, whose 'Dog: Behaviour, Evolution, Cognition' is also reiewed on infopet.


If Dogs Could Talk: Exploring the Canine Mind

Vilmos Csanyi 

Review by Diana Attwood

This book seems to be very much written in different parts, some of which I found made for fascinating , inspiring and very enjoyable reading, some of which I found quite difficult to really see any great significance in, and some of which made me wonder that the author really knows very much about dogs!

Vilmos Csanyi begins this book with an overview of his theories of how man and dog came together , which a very brief reference to Coppinger’s theory of how dog evolved from the wolves that scavenged around human settlements, and talks in more detail about comparisons between our individual species characteristics. From there he goes on to look at how many of these very similar characteristics seem to have evolved in dogs directly from their close relationship with humans as many of these characteristics are not typical of their direct ancestor, the grey wolf, but are in various naive forms often found in man’s direct ancestry, among various species of ape.

I quite liked the idea here that man may have created dog in his own likeness!

As a scientist, he talks about the various trends that scientists have followed over the years and how science has traditionally dismissed ideas of emotion and higher intelligence in animals because science is purely about that which can be proved and written down as fact, and makes note of the limitations with animals in that you cannot compare an animal’s wants with a humans wants because ultimately both will have very different wants for different reasons. An example he puts forward – can a dog recognise itself in a mirror? Why should it want to? (I questioned why should we presume that they can physically see themselves that well when their strength is generally in their sense of smell and their eyesight is more adapted towards the moving object rather than the still? A dog might smell the mirror but they can’t smell their reflection).

As a dog lover, Vilmos Csanyi talks about his dogs with a huge amount of affection and clearly loves the companionship of his dogs. His dogs are purely pets and as pets he looks at how they interact and communicate with other members of his household and cites many examples of their ability to both understand social rituals and politenesses, and communications.

This part of the book, I’m afraid (and sorry Mr. Csanyi), slightly irritated me. Maybe a little was lost in translation, maybe some was lost in cultural differences, but many of his own observations seemed to be bias towards his own dogs abilities and drives, which he has taken down to be typical examples of how all dogs minds might work. An example he pointed out regarding his dogs natural understanding of the rules of the house were that although he has many things that could be knocked over in his home, his dogs have never done so.

In one of his books, the late John Fisher looked at his two dogs at the time, a small terrier and a weimaraner, and noted that the differences in size meant that while one could squeeze through as small gap, he might hanker for the opportunity to get to that gap and take himself off for a walk, where the other would not be bothered about the gap, because he couldn’t get through it. The same gap that held great significance for one, had none for the other. The larger dog however, had a much keener eye on the worktops and dining table, as his height meant that this was accessible to him and so something he sought to gain access to.

My Cindy is a weimaraner. A long, tall dog selectively bred over centuries to hunt and seek out prey with its powerful sense of smell. Atomically, her long ears and flues developed as means of giving her an even greater draw of scent than say a prick eared dog with a tight muzzle, or a tulip eared dog with a short muzzle. She frequently knocks things over in the home, because, I believe, for the most part at close distances, she ‘sees’ with her nose rather than her eyes, and while her nose is quite incredible for picking up and telling her a great deal about warm bodied prey objects, it’s not that good at working out the precise dimensions of static objects of furniture.

He also devotes some of his book to the understanding of different breed types. It's clear throughout that Csanyi has a preference for mixed breed dogs over purebred dogs, but his generalisations sometimes seem desperately unfounded and show absolutely no appreciation of dogs as a working animal. He does have notes and observations from dogs trained to work closely as guide dogs, but this is a relatively new role for a dog that has been selectively bred for centuries to pick up and retrieve shot game – and he completely neglects this aspect of all dogs - the animal that man created in his own likeness, not to be a pet but to work alongside him.

Another example showing how he misses insights from knowedge of working breeds is that he uses the act of pointing, in that he might point in a particular direction and observes that his pet dogs understand this gesture – to his fellow countrymen who work in the field with the Hungarian Vizsla, this would most certainly perplex them – of course they do, they’re pointers and will even follow each other’s points!

One of my Weimaraner’s favourite games is the pointing game. She loves to fly across the field in the direction I point to, and fly back as I then call out and point in another as if we are looking for some mystery invisible object.

And at this point, I really become confused by the whole point of the book. He talks about all the many ways in which our dogs communicate with us, and how we can read and understand this communication, but then goes on to suggest we breed to create a dog of higher intelligence that can actually speak. What is the great need for speech when you can communicate perfectly well with a point, and ultimately what’s the point of talking if no one is listening. I’m sure Flip and Jerry might answer ‘...but we *are* talking Vilmos, what more do you want?’

Csanyi does seem to overrate the spoken (human) word, and link it too strongly with intelligence. Research by linguists, mentioned by Steven Pinker, shows that intelligence and language are separate, ie one can have people who talk grammatically and fluently and yet have very poor reasoning power, and, on the other hand, people who have lost language through a stroke or other reasons, and still have good reasoning power.

I think that there are many interesting ideas in this book, but there are also some major gaps.  Vilmos might do well to look more closely at working dogs if he wants to achieve a fuller understanding of how dogs communicate with humans. Viszlas might make for a particularly interesting study for him as they are native to his country and have been bred to use their own minds and communicate with their owners in the practice of the role that they were specifically developed for.