Endings have the richest critical literature of any subject in this book. Five strands are used here, each because it answers a question a writer meets in revision; the Reference treatment adds Torgovnick’s typology of closural relations, the cognitive evidence in full and the history of the ending period by period.
Why does a reader who knows endings are artificial still want one?
Frank Kermode’s The Sense of an Ending (1967) starts from the observation that we live in the middest, born after the beginning and dying before the end, and make sense of the middle by positing ends. He calls the resulting shapes fictions of concord, and his model is the clock: it makes identical sounds, and we hear tick-tock, a beginning and an end with an interval that the pair organise into a duration. Nothing in the acoustics licenses the distinction; the listener supplies it. Kermode’s point for a novelist is that sophisticated readers know an ending is imposed and want it anyway. Endings are discredited but not abandoned, which is why a novel that refuses shape on principle often disappoints readers who endorse the principle. From Aristotle he takes peripeteia, the reversal that falsifies an expectation and satisfies a better one, and the craft form of it is the rule this chapter uses throughout: an ending should be inevitable in retrospect and unforeseen in prospect.
What makes a last line feel final?
Barbara Herrnstein Smith’s Poetic Closure (1968) separates termination, a text simply stopping, from closure, the reader’s sense that the stopping is the appropriate conclusion of the structure. Because a poem is short enough to analyse exhaustively, she could describe closural devices with a precision that transfers to prose almost intact. A closural allusion refers, at a terminal position, to something terminal (sleep, night, snow, a departing train), and is the cheapest device in the language, which is why so many endings feel right and audit badly. A terminal modification alters a pattern at its last instance, so that in a novel the last chapter that does not do what the previous forty did is read as closure even by readers who cannot name the pattern. An unqualified assertion, a short declarative sentence without hedging, closes by grammar alone. Her account of modern poetry adds a principle the chapter relies on: an open ending is a closural strategy, not the absence of one. It works by withholding devices the reader knows are available, so it depends on the reader’s competence in the conventions it declines, and reads as integrity only in a book that has taught the reader to expect it.
Why does a weak ending damage the whole book?
Peter Brooks’s Reading for the Plot (1984) describes reading as anticipation of retrospection: we read forward in expectation of a moment from which the whole will become legible, and tolerate the middle because we expect an end that will say what it was. Meaning is conferred from the terminus, which is why an ending can make earlier material better without changing a word of it, and why a reader who has hoped for four hundred pages that two people will marry feels cheated by a marriage the book has not earned. Readers want the ending the book requires, not the one they wanted. Peter Rabinowitz’s Before Reading (1987) supplies the interpretive half: readers follow tacit rules of notice, and the last sentence holds a privileged position, read as a statement about the book whether or not it was written as one.
The cognitive evidence is suggestive and limited. Donald Redelmeier and Daniel Kahneman (1996) found that patients’ memory of a painful procedure was predicted largely by its worst moment and its last, the peak–end rule. The studies concern brief physical discomfort judged soon afterwards; extending them to a novel read over weeks is an analogy, not a finding, and the retroactive lens rests on the literary evidence without them.
What does “earned” mean?
Aristotle’s Poetics requires the incidents of a plot to follow one another by necessity or probability, because and not merely after, which is the test Scene and Sequel (Chapter 2) applies to every causal chain. Two refinements carry into revision. Probable is enough: an ending need not be the only possible outcome, only one the established people in the established world would plausibly produce. And the objection to the deus ex machina is to externality, not improbability: an outcome produced by something outside the causal system the play has built. Aristotle was arguing against a living practice; the god arriving at the end was Euripides’ signature, and later writers made its arbitrariness the point.
What if closure is the problem?
D. A. Miller’s Narrative and Its Discontents (1981) denies the assumption behind everything above. Narrative consists of the narratable, instability, lack and desire, and closure is the elimination of the instability, so closure ends the novel’s own condition of existence. He finds traditional endings thinner than the books they end, and written in a different register, summary instead of scene. Rachel Blau DuPlessis’s Writing Beyond the Ending (1985) adds a history: the nineteenth-century novel offered a heroine two endings, marriage or death, so a closural convention can encode what a life is allowed to be for, and a mismatched last chapter can be the sound of material forced into an unavailable shape. Things Fall Apart ends by showing the form that would have contained it, the District Commissioner reducing Okonkwo to a paragraph.
This chapter accepts Miller’s diagnosis and disputes his prescription. Closure does end the narratable, and every ending is a loss; but a novel is also an object with a shape, and the shape is available only from the end. The craft question is what the reader is given in exchange. The account predicts that strong endings close the book while leaving one thing moving, and the readings in Worked Analyses bear it out.
A caution about period
Most prescriptions here are calibrated to Anglo-American practice since about 1960. In that tradition the unresolved close has become the house style of literary fiction, while crime, romance and thrillers keep strong resolution conventions, so an open ending no longer signals a deviation and must do the work any other choice does. Victorian wind-up chapters and Tolstoy’s eighth part show that the short aftermath is a period preference. Henry James named the underlying problem in his preface to Roderick Hudson (1907): relations stop nowhere, and the exquisite problem of the artist is eternally but to draw, by a geometry of his own, the circle within which they shall happily appear to do so. The ending is that circle, and the word happily sets its standard: the appearance must satisfy, not merely be asserted.
See also: the Reference history of the ending, from the Victorian settlement to the present, with DuPlessis and Miller at full length.