Karen Armstrong could not be accused of failing to put her cards on the table.

By the end of her first paragraph, she has asserted baldly that "modern society has made a scapegoat of faith". She is taking issue with a cliché, the routine claim that religion, advertising itself as humanity's finest expression, has been responsible for most of the woes of the species.

The charge, like attempts to refute the charge, has probably been around since chieftains and kings first began to make an organised business of killing. Few people, believers or otherwise, have failed to notice the contradiction between religions which outlaw murder yet seem to justify, sometimes to sanctify, mass slaughter. The language of Christians is blunt: "Thou shalt not kill." Yet there we are, century after century, marching off "with God on our side".

As Armstrong writes, "eerily" the charge is made "in the same way almost every time: 'Religion has been the cause of all the major wars in history.' I have heard this sentence recited like a mantra by American commentators and psychiatrists, London taxi drivers and Oxford academics." There is reason enough for that, you might think, but it also happens to be untrue. Religion did not cause the two great global conflicts of the 20th century. Whatever the trappings, religion's role in other wars, Armstrong argues, has been less clear cut than the cliché would have us believe.

Though a distinguished writer on religion, her interests in this are more than simply academic.

In a long survey, beginning with the first faltering attempts to create agrarian societies in the Levant, there is an underlying argument. Or rather, the old argument between religion and what are termed Enlightenment values is continued.

As Armstrong puts it, "so indelible is the aggressive image of religious faith in our secular consciousness that we routinely load the violent sins of the 20th century on to the back of 'religion' and drive it out into the political wilderness."

The claim that the world has become secular is as familiar as it is odd. It has become a preoccupation of the faithful in Europe and - above all - the United States. Secularism, so called, is depicted as conducting a warfare of its own. But as Armstrong knows and explains, the argument involves a very narrow view of both the world and religion. Set a few noisy (and unoriginal) atheists against billions whose lives are permeated by faith and you wonder about all the fretting.

In any case, the basic charge - that endless wars have been fought for the sake of gods - remains intact. What is it about humanity that allows it to receive and accept messages of love and compassion yet use faith in such precepts to justify mass murder? Freudian hokum invoking the trinity of id, ego and superego might once have been called upon. Armstrong prefers neuro-anatomy and a near-Marxist account of elites, class, oppression and exploitation.

"Each of us," she writes, "has not one but three brains which co-exist uneasily." We acquired this collection at various stages of evolution. The oldest is a reptilian remnant, utterly self-interested; the second, the limbic system, allows empathy; the youngest, acquired perhaps 20,000 years ago as the neocortex or "new brain", grants us self-awareness. We can stand back, as Armstrong puts it, from our primitive instincts. We can also be imbued with faith and utterly murderous.

This is neat. It is also, for the sake of an argument, convenient. Armstrong understands religion in terms of the human search for - indeed, need for - meaning. Most of the major faiths make larger claims to do with a deity and absolute, eternal truth. If that is what is going on, surely the ugly, insistent claims of our buried "old brain" would be overwhelmed. It might be a mistake to claim, crudely, that religion causes war. But to put the question in the context of a fine and eloquent book: why has the impulse to faith failed to suppress our brutish taste for warfare?

Armstrong deploys a socio-economic argument. Early humans subsisted, hunting and gathering, never achieving the food surpluses on which warfare depends. Co-operation was essential; hierarchies pointless. Social organisation must have been near-communist, as the author suggests.

Only the coming of agrarian societies capable of sustaining large populations and elites made war possible. In essence, the species had something worth stealing and a peasantry worth exploiting.

The analysis is hard to dispute. When was it ever otherwise? What Armstrong calls the "structural violence" of society has been evident since the first cities - and their defensive walls - began to appear in the Middle East. A tiny minority, with their armies and their priests, imposed themselves on the majority. The masses suffered, time and again, but civilisation and its arts, requiring the leisure available only to those who did no work, flourished mightily.

The picture is bleak, but certainly accurate. It is part of Armstrong's task to explain that while the religious inveighed often enough against violence and greed they failed, as it were, to beat the system. The impulse to faith might be ancient, even natural in a "meaning-seeking" species, but organised religions, especially state religions, have a tendency to defeat their own dreams, the dream of a just society in particular.

Armstrong does not claim to examine every instance of warfare in a god's name, but she has a good try. The Crusades, the Inquisition, the Wars of Religion, even modern "jihadi" terrorism: each is investigated in a book that is long and yet, fluent and elegant, never quite long enough. Fields Of Blood is as much about the nature of warfare as it is about faith.

What soldiering does to people and societies is given acute consideration. As Armstrong explains - and as most of us probably grasp - individuals are changed in strange ways by war. Even amid the horror, perhaps because of the horror, many become more fully aware and, as has been reported down the ages, more completely alive. It is another unpalatable fact for which neuro-anatomy does not seem to offer a complete explanation.

"It is simply not true that 'religion' is always aggressive," writes Armstrong in her concluding pages. A few paragraphs later she concedes, however, that no state "in history has... not incurred the taint of the warrior". If a species could be said to have a personality, these two contradictory traits exist side by side in us. The author hopes that in a modern, interconnected world the paradox can be resolved, that we each have a responsibility, whether religious or secular, to work towards that end. History does not hold out much promise.

Armstrong does not dismiss what she calls secularism; quite the reverse. She accepts that faith traditions fail to prosper when they become entangled with the business of government. Equally, her view of religion is broad and deep, identifying it both with a universal need for meaning, and with the human desire for community. Exploitation and oppression continue regardless, as she knows, but these provide a challenge for the godly and the godless alike. The proposition, like the book, is noble.