Should the proverbial extraterrestrial include earth on his intergalactic tourist pass this summer, then there are worse places he could look for guidance on native customs than the Natural History Museum in London. Shuffling past the rows of plasticky dinosaurs and petrified mammals, he probably wouldn't end up much wiser about the creatures we share this earth with. But he would find out a great deal about Homo sapiens: our needs and our blindness, our intellectual fashions, our lonely desire to taxonomise life and, by naming it, push it further away.

John Bradshaw's new book, In Defence of Dogs, unwittingly does the same thing. By giving the reader an overview of mankind's relationship with both dogs and wolves, he also shows us ourselves  our need for connection but our insistence that it be on our terms, our hellish good intentions.

None of this is intentional. As a biologist and founder of the anthrozoology department at the University of Bristol, Bradshaw's aim is to prove that the dog-training model of the past few decades is flawed, based as it is on the premise that dogs are basically wolves with nicer table manners. Since dogs share 99.96% of their DNA with the grey wolf, it's been a persuasive idea. Unfortunately, in order for it or any other training model to work, it depends on us having a complete 360-degree understanding of both dogs and wolves. As Bradshaw shows, we have neither.

Because we are human and change our minds, the way we think of animals changes too. Until very recently, wolves were our demons, the creatures of antique nightmare ready to rip the throat out of civilisation. Their packs were ruled by despotic alpha pairings who destroyed any sign of difference or dissent, and wolves were meant to spend their lives in an everlasting struggle for dominion over their environment and one another. And, since we saw them as properly, biblically bad, we granted ourselves full licence to kill them wherever possible. Which in turn meant we reduced the global gene pool to only the smartest, wildest wolves  the ones as wary of humans and as distinct from dogs as it was possible to get.

Wolves were our Hyde, so dogs were our Jekyll. Leaving aside the question of their genetic origins, it didn't take us long to realise that dogs had qualities we could use. The intelligent sorts could be used to herd sheep or guide the blind, the aggressive types could guard our families, and the affectionate ones could be both loyal companions and handy kitchen dustbins. But if every dog from meanest rottweiler to prissiest pomeranian was really a wolf beneath the skin, it naturally followed that they would try to dominate their owners at the first sign of relaxation. One glimpse of a warm sofa cushion and your chihuahua would start calling for fine wines and announcing itself head of the household.

Bradshaw's contention is that the dog-as-wolf idea doesn't work. In recent years, we've realised that wolves don't just re-enact an endless lupine version of Lord of the Flies, and that just because they like the occasional lamb cutlet doesn't mean they're all monsters. Meanwhile, all those thousands of years of domestication have made dogs uniquely capable of loving both people and other dogs, and uniquely tractable. As Eddie Izzard once pointed out, had Ivan Pavlov tried his famous experiment in classical conditioning on cats, he would have died penniless and insane.

Despite plenty of strong and contentious material, In Defence of Dogs isn't the easiest read. It's a biologist's book designed as part of an ongoing conversation between other academics, and though Bradshaw's style is supple and fluent, he has a tendency to err on the side of too much science and not enough story.

It also takes him until the last chapter to discuss the differences between breeds of dog. Until then, dogs are just dogs, as same-ish in habits and behaviour as goats or blackbirds. But, as even non-dog-owners could tell you, dogs, like people, are individuals. There might be particular traits shared by one or other breed, but there are stupid collies and clever setters, timid ridgebacks and solitary labradors. Some sheepdogs are scared of sheep; some guard dogs are only useful as speed bumps. As with humans, there might be a few overriding organisational principles  food, love, shelter, warmth  but one of the reasons that the bond between dogs and people has sealed so strongly is because they, like wolves and like us, have their own souls. To see most expressions of character in dogs as mere anthropomorphism is, once again, to have miscommunicated. It is precisely those differences in personality that the Inuit once harnessed among their sled dogs  intelligent leaders, excitable pullers, stolid team types  and which zoologists are also beginning to recognise within wolf-packs.

The trick, as Bradshaw points out, is to allow dogs to be dogs. In his view, that probably means preventing them from breeding with their own choice of mate and moving away from working types towards the more amenable companion types. That way, he believes we might have a chance to compensate for the tens of thousands of dogs put down every year, and, equally urgently, for the incestuous cruelties of pedigree breeding.

Again, it's all about what we do to the dog  how we rear them, how we think they ought to be. But communication is supposed to be a two-way process, and one of the most enriching things about owning a dog is finding that they train you back. You may think it's all about throwing sticks and fortifying walks, but you very quickly realise that what you're receiving is an education in a whole new sensory world, in a life not dominated by the mind, and in the ordinary astonishment of what's going on right now.

Somewhere between Bradshaw's Resource Holding Potentials and Thorndike's Puzzle Boxes is the tale of two species groping towards a common language. In his plea for a broader and more generous understanding of dogs are concealed all our centuries of hope, science, love and revisionism, and what he gives us is not so much a defence of dogs, but a portrait  flaws and all  of ourselves. In the end, it's sometimes difficult not to wonder which we've treated worse: our oldest enemy, or our best friends.

Bella Bathurst's most recent book is The Bicycle Book (HarperPress)

If you were a dog just over 100 years ago, life would have been simple. You would likely have been gainfully employed  perhaps hunting, herding or guarding  and provided you did your job, your owners would have accepted that you were sometimes messy, loud or unpredictable. Most dogs today are never expected to work, even though they are often still tuned into functions their breed has fulfilled for thousands of years. Instead, they are expected to behave like small children, yet be as independent as adults. To make things worse, our culture is awash with myths that prevent dogs being properly understood  in particular, the enduring idea that they harbour a powerful desire to dominate their family pack. Put simply: dogs are on the brink of a crisis. And as we have put them there, it is our responsibility to help them.

This is the thesis of John Bradshaw's scholarly yet passionate book In Defence of Dogs, which is nothing less than a manifesto for a new understanding of our canine friends. It is an attempt to "stand up for dogdom"  that is, dogs as they truly are, not as we assume they are. As a canine expert and dog-lover, Bradshaw is dismayed that our treatment of dogs is based on so many mistaken beliefs and assumptions. He wants to set the record straight now because canine science has made huge advances in recent decades.

He starts by demolishing the notion that dogs are essentially aggressive creatures seeking dominance, which is based on discredited research into wolf packs. It is now known that wolves  the direct ancestors of dogs  actually live in harmonious family groups. Packs are not dominated by "alpha wolves", but are fundamentally cooperative. Bradshaw is determined that the "dominance theory" be banished. But while enlightened trainers and owners have got the message, many more still subscribe to techniques aimed at ingraining fear and subservience into dogs. For Bradshaw, these are not only misguided and cruel, but joyless.

His account of the evolution of dogs is fascinating. Surveying the latest research, he concludes that the dog's epic journey towards domestication probably started around 20,000 years ago. Dogs have become almost a separate species from wolves, and their evolution continues to confound biologists. What Bradshaw is keen to stress, though, is the unique evolutionary pact between humans and dogs: we have programmed into them a deep need for relationships with humans, which we must treat with respect.

This material underpins Bradshaw's most compelling chapters, which explore the emotional lives of dogs. The revelation here for many dog owners might perhaps be that dogs' emotional repertoires are much more limited than we generally think. Research confirms that most dog owners are convinced their dogs can feel and display complex emotions  particularly guilt. In fact, there is almost no evidence for this; dogs simply do not have the self-awareness for such emotions. But in persisting with the notion that dogs have this advanced understanding of their actions  and our expectations  we end up punishing them in ways they cannot understand. Dogs are specialists in love, fear and joy. But we must stop assuming their knowledge of emotions beyond their grasp.

Elsewhere in these sections, Bradshaw tackles the question: "Does your dog love you?" The answer is yes: probably even more than you think. Dogs are profoundly attached to their owners, and this love  a term Bradshaw happily uses  is often at the root of their apparent misbehaviour. For example, dogs not properly trained to understand that when we leave we will return can be plunged into the depths of anxiety when we are not around. Bradshaw estimates that up to 20% of dogs suffer from "separation distress" when left alone at home.

Most people can probably intuit that human progress has cut many dogs off from the activities that previously gave their lives meaning. (Anyone who has spent time with a border collie will know that their boundless desire to herd everything from pushchairs to small children betokens something of a behavioural hangover.) And Bradshaw's arguments against pedigree breeding play into an existing public debate (breeds heading for extinction due to the demand for perfection).

His sober argument finds an unlikely echo in Jan Bondeson's slightly bewildering volume Amazing Dogs (Amberley Publishing, £20). An eccentric romp through canine history, it nevertheless shares the same thesis: dogs are poorly served by our misunderstanding of them. This is made clear in his chapters on the glum history of "canine intellectuals", who wowed 19th-century crowds around Europe with their supposed skills  from poetry to arithmetic to clairvoyance. Elsewhere he celebrates the true over-achievers from canine history: Shakespearean actors, charity-collectors, and dogs whose loyalty resulted in years-long graveside vigils for their dead masters.