Three questions. Is it possible to re-establish the lost bond between literature, history and the world, while still maintaining a full sense of the irreducible singularity of literary texts? Second, can literature itself be conceived as a world? And if so, might an exploration of its territory help us to answer question number one?

Put differently: is it possible to find the conceptual means with which to oppose the central postulate of internal, text-based literary criticism—the total rupture between text and world? Can we propose any theoretical and practical tools that could combat the governing principle of the autonomy of the text, or the alleged independence of the linguistic sphere? To date, the answers given to this crucial question, from postcolonial theory among others, seem to me to have established only a limited connection between the two supposedly incommensurate domains. Post-colonialism posits a direct link between literature and history, one that is exclusively political. From this, it moves to an external criticism that runs the risk of reducing the literary to the political, imposing a series of annexations or short-circuits, and often passing in silence over the actual aesthetic, formal or stylistic characteristics that actually ‘make’ literature.

I want to propose a hypothesis that would move beyond this division between internal and external criticism. Let us say that a mediating space exists between literature and the world: a parallel territory, relatively autonomous from the political domain, and dedicated as a result to questions, debates, inventions of a specifically literary nature. Here, struggles of all sorts—political, social, national, gender, ethnic—come to be refracted, diluted, deformed or transformed according to a literary logic, and in literary forms. Working from this hypothesis, while trying to envisage all its theoretical and practical consequences, should permit us to set out on a course of criticism that would be both internal and external; in other words, a criticism that could give a unified account of, say, the evolution of poetic forms, or the aesthetics of the novel, and their connection to the political, economic and social world—including telling us how, by a very long (indeed historical) process, the link gets broken in the most autonomous regions of this space.

So: another world, whose divisions and frontiers are relatively independent of political and linguistic borders. And with its own laws, its own history, its specific revolts and revolutions; a market where non-market values are traded, within a non-economic economy; and measured, as we shall see, by an aesthetic scale of time. This World of Letters functions invisibly for the most part, save to those most distant from its great centres or most deprived of its resources, who can see more clearly than others the forms of violence and domination that operate within it.

Let us call this mediating area the ‘world literary space’. It is no more than a tool that should be tested by concrete research, an instrument that might provide an account of the logic and history of literature, without falling into the trap of total autonomy. It is also a ‘hypothetical model’ in Chomsky’s sense—a body of statements whose working out (if risky) may itself help to formulate the object of description; that is, an internally coherent set of propositions.footnote1 Working from a model should permit a certain freedom from the immediate ‘given’. It should, on the contrary, allow us to construct every case afresh; and to show with each one that it does not exist in isolation, but is a particular instance of the possible, an element in a group or family, which we could not have seen without having previously formulated an abstract model of all possibilities.