The length and complexity of India's civilisation are such that one-volume histories of the subcontinent are rare and single-author histories are rarer still. The late Burton Stein was one of the few historians of India to carry out research on widely differing periods of that civilisation and his last book, which was virtually complete by the time of his death, is marked by a corresponding breadth of vision and originality.

In his opening chapter, Stein expounds two concepts that shape his analysis. The first is the idea of the segmented state, borrowed from African anthropology, which allows community leaders to have real political and administrative power while recognising the sovereignty of a single king. Stein sees communities not simply as groupings of caste or religion but as people with shared rights in particular environments. In the book, this allows the development of socioeconomic themes such as rural stratification, the growth of trade and merchant classes and the successive stages of urbanisation.

The second concept is one of the changing relationship between communities and states. The earliest phase of Indian society was one in which communities existed without states (Stein here follows a revisionist interpretation of the Harappan civilisation). The next (from 800 BC to AD 300) was one in which communities operated as states, a phase best exemplified by the so-called republics of the Gangetic plain and not, in Stein's view, contradicted by the superficial centralisation of Ashoka. The third phase, running from the Gupta dynasty to about 1700, was one of the coexistence of communities and states, a process that can be illustrated, for example, in the relationship between local institutions of various kinds and the Mughal regime. By the 18th century, Stein believes, communities had become "decorticated" and survived from then on as ideological constructs, while social and political relations operated in a class framework.

One advantage of this approach is that it allows the history of the peninsula to be seen as part of the story of the subcontinent as a whole instead of being merely an appendix to the history of the northern empires. Disappointingly, though, Stein misses the opportunity to apply his insights to some parts of the north. For example, the five centuries of confusion that lie between the death of Harsha and the rise of the Delhi Sultanate are passed over in a few lines, even though it would seem that Stein has at last provided us with the analytical framework to make sense of them. Indeed, the whole book would have benefited from a closer relationship between the interpretative introduction and the more descriptive chapters that follow, as well as a greater emphasis on the ancient and medieval periods. By the book's half-way point, Stein is already into the period of the East India Company and "decorticated" communities. His last chapter outlines the post-independence history of India and Pakistan.

ÌÇÐÄVlog

There are other places, too, where the book's coverage is less complete than some might wish it to be: for example, in aspects of Sikh history or on contemporary trends in Indian politics. But a book of this size and scope has to be selective in its details.

What impresses is Stein's ability to take the long view of Indian history and to develop his analysis across so many centuries and through such diverse material. It is also a mark of his achievement that his book stimulates discussion of matters that he has touched on only fleetingly himself, for it is an argument as much as an account of events. It is also genuinely a history of the peoples of India, not just of their dynasties and politicians. Stein gives careful consideration to religious developments, to economic and cultural links with the outside world, to environmental issues and to the changing position of women. His ironic account of the Buddha's struggling to deny salvation to women will be cherished by many.

ÌÇÐÄVlog

The book has excellent maps and its chronological tables are helpful, but the suggestions for further reading are inadequate. Although a full bibliography would clearly be out of the question, the reader should at least be guided around the controversies that are mentioned in the text.

Richard Newman is a research associate, School of Oriental and African Studies, London.