One of the functions of literary fiction, in particular narrative fiction, is the construction, circulation and maintenance of world-models.  Literature, Lotman taught us, is a secondary modeling system: using the primary modeling system of language as its vehicle, it constructs models of and models for reality.  It also models itself: narrative fictions regularly embed within their own continuums secondary worlds – inset narratives, found manuscripts, ekphrastic descriptions, remediations of non-verbal media, micro-worlds and paraspaces, etc. – that mirror the primary worlds framing them.  Such structures en abyme have typically been treated as uncanny disruptions, fatally compromising fiction’s world-modeling function, at worst summoning up the specters of crippling paradox and infinite regress.  In fact, world-modeling and self-modeling are interdependent functions of fiction; they complement and sustain each other.  Internal scale-models make the “outer” fiction’s model of the world salient.  Far from disrupting the primary world, they hold a mirror up to it, providing the reader with a kind of schematic diagram of it, or an instruction manual for its proper operation. Moreover, the relationship between the “outer” world and the internal scale-model, the way one maps onto the other, can itself serve as a model for the relationship between the fictional world as a whole and the real world – the world “out there,” beyond the text.  So internal scale-models yield knowledge of the fictional world, but also of how the fictional world models the real.

          The paper revisits the literature on mise en abyme, as well as Jameson’s powerful notion, derived from the urban planning literature, of cognitive mapping.  Case-studies include Cervantes’ Don Quixote, read in the light of Fernand Braudel’s cognitive mapping of the Mediterranean world in the age of Philip II; micro-worlds and scale-models en abyme in science fiction (Gibson, Sterling), and in the American mega-novel (Sorrentino, Pynchon, Barth).