BOOK REVIEW Literary approach to a complex relationship Edward W. Said.Culture and Imperialism, Vintage, London, 1994 , pp. xxxii+444. The questions concerning the relationship between culture and imperialism that Said has posed in his most recent book are important and complex. They are ones that many of us have faced in thinking about imperialism. Said's own attempt to answer these questions is , I think, deeply flawed, but it is rich with ideas. As a beginning, his writing may help us think through the problems in an alternative way. In this very brief review, I'll indicate what I see to be a few of the fundamental problems with his analysis. Said's basic approach into the vast topic of culture and imperialism is to review British and French novels of the nineteenth to mid-twentieth century. Said considers novels to be 'immensely important in the formation of imperial attitudes, references and experiences' (p. xii). Novels may indeed be 'particularly interesting to study' (p. xii) but the reader is not given any evidence that novels were important at a mass level in Britain and France. How many people actually read lengthy novels in the largely illiterate European countries of the nineteenth century? Novels should certainly be studied but their significance as determinants or representations of empire should not be inflated. In following Said through his literary approach to imperialism, the reader may well forget what is missing: a study of popular culture and everyday practice, whether in Europe or in the colonies. Said has made an advance over his book Orientalism (1978) by intertwining a reading of European writers with writers of the colonies. He is now willing to admit that he made a serious error in omitting the voices of the colonised in his previous book. In C&J, Said places the European texts side by side with those from colonies. For example, Camus's novels, in which France's possessions in North Africa play a significant part, are discussed in conjunction with Fanon's writings on the Algerian struggle for independence. This methodology produces much of what is valuable in the book and is a clear departure from the Social Scientist, Vol. 23, Nos. 1-3, January-March 1995