Contra Susan Sontag and a whole generation of structuralist literary critics, Kermode might have titled these essays--the Norton Lectures on Poetry, 1977-78--as "For Interpretation." Because all narratives share a "radiant obscurity," as we read we honor this mystery by helplessly trying to figure it out. Hermeneutics is usually the province of biblical scholarship, so Kermode resolves to start right there. Laminated with "secret texts," "midrashim," and corollaries to the Old Testament, the Gospels are a perfect ur-text: agents can be seen to become characters in the course of successive interpretation. Mark, the earliest written gospel, is a harsh story, purposely elusive, almost taunting. Matthew becomes more vivid, but also lops off edges that can make the reader/believer very edgy. Luke and John add verisimilitude--novelistic touches, necessary alignments. The synoptic Gospels, therefore, are created, Kermode argues, like any other text: they receive and consolidate sketchy mysteries, respond to the historical realities of their time (and prospective audience), and in their structures behave like any fiction: the how of the telling shapes the narrative fully as much as what's being told. Not to interpret, Kermode's argument goes, is to write off this hermetic, layered dignity of texts, to fix them to an ideology, to deny their mystery, treating them either as neutral architecture or journalistic propaganda. (There are modern references also--to Pynchon, Green, Kafka.) Though Kermode slips into jargon now and then, the thesis is well wrought, the scholarship varied and well-distributed, and the examples clear and deft.