I found this book fascinating, but, since it concerns a subject I know little about, I have a hard time evaluating its reliability. The initial idea, which I have seen many times before, is that our brains are divided into three main parts: the "reptilian brain" (basic functions), the "mammal brain" (emotions), and the "human brain" (abstract thought). Love is a function of the "mammal brain", the limbic system, and is essential to the formation of relationships, in particular parental relationships. Hence, only mammals can have relationships; and, since the formation of relationships relies on neural hardware that we share with other mammals, it is perfectly reasonable to talk about people having genuine emotional relationships with non-human mammals like dogs, horses and cats. You cannot, however, have a relationship with a reptile.
The authors, all psychotherapists, develop this idea by arguing that emotional relationships mediated by the limbic system are far more important than many people today wish to acknowledge, and that this lack of understanding is causing us untold damage. They start by looking at small babies, who over and over have been shown to be absolutely dependent on close physical contact with their mother or a substitute. The reason, it is claimed, is that an infant's physiology is an open system. He is not capable of regulating his bodily functions on his own, but needs the mother's presence to help him do it. If he is separated from his mother, he first becomes stressed and agitated, and if the absence is prolonged lapses into an inert, uncaring state. There are many directly verifiable physiological correlates (changes in heart rhythm, hormone balance, immune system responses, etc), which in babies can easily end up being fatal. Adolescents and adults who are separated from loved ones react in a similar, but less extreme way, but the basic argument is that the pain of loss is not an illusion. Heartbreak hurts for very real, physical reasons, that come from our basic mammalian biology.
The book then looks at memory, in particular unconscious learning processes; it says these are also mediated through the limbic system, and that learning happens by acquiring behavioral patterns from the people with whom we have close emotional relationships. In particular, children generally learn their emotional responses from their mothers. Children whose mothers are not close to them, or who use an inconsistent parenting strategy, become adults who will have life-long emotional problems. Therapy, the authors say, works by establishing a close emotional relationship between the patient and the therapist, similar to the one the patient had with their parent. The success or failure of the treatment does not depend on which psychological school the therapist belongs to (they have a hearty contempt for Freudian theory). Rather, it depends on what kind of person the therapist is, since the patient is essentially trying to become the therapist at an emotional level.
I have trouble deciding whether to believe all this. I do think they have a good point about how modern Western society underestimates the importance of emotion. One interesting section offers an explanation of why "alternative medicine" has become so popular. Even if the treatments are scientific nonsense, patients don't just go to the doctor to get a scientific diagnosis, but because the emotional relationship with the doctor is important to them. In the US, doctors have become so distant and impersonal that they no longer fulfill this function; a homeotherapist may be handing out quack medicine, but usually he also offers warm, human contact. The authors say this can in many cases be nearly as useful, and studies do indeed show that the placebo effect is very strong.
It's a thought-provoking book; they clearly believe they're telling us something important, and they've based it on a lot of reading and clinical experience. Even if some of it is no doubt wrong, my feeling is that enough of it is right that it's well worth reading.