The theory below is limited to what the chapter’s tools use. Each strand opens with the question a writer in revision is asking, and then gives the source that answers it.
How does a reader know that a character has changed?
The reader knows by watching choices. Aristotle’s Poetics, in S. H. Butcher’s translation, puts it in one sentence: “Character is that which reveals moral purpose, showing what kind of things a man chooses or avoids.” The Greek words are ēthos, character, and prohairesis, deliberate choice, and the claim is that the first is visible only through the second. Aristotle ranked plot above character, and the ranking follows from the definition: a character is known by what they choose, so the choices, which are plot, come first.
Robert McKee’s Story (1997), a widely used screenwriting manual, restates the point as a working rule: “True CHARACTER is revealed in the choices a human being makes under pressure – the greater the pressure, the deeper the revelation, the truer the choice to the character’s essential nature.” McKee separates true character from characterisation, the observable sum of a person’s traits, and the separation is useful in revision. A writer who wants a character to seem changed can alter the characterisation, a new haircut, a softer voice, and nothing beneath it moves. A change in character is a change in what the person chooses when choosing costs something, and the more it costs, the more the reader believes it. The household draft in The Problem changes Elena’s understanding and leaves her choices as they were, which is why the reader does not believe it.
Does every character need to change?
No, and the clearest early account says why. E. M. Forster’s Aspects of the Novel (1927) divides characters into flat characters, “constructed round a single idea or quality”, and round characters, whose test is that they are “capable of surprising in a convincing way”. Forster defends flat characters, which are easy to recognise and remember, and holds only that round ones are needed to carry tragedy for any length of time. His test of roundness is close to a description of the hinge. A surprising decision is one the reader did not predict from the character’s earlier choices; a convincing one is a decision the reader, looking back, finds prepared. A change that is predicted fails the first condition and reads as a schedule, and a change that is unprepared fails the second and reads as the author’s convenience.
Forster’s division also licenses the steadfast and refused arcs in The Taxonomy. A round character can refuse to change, and the refusal can surprise; a flat character can change other people without changing at all.
Where do the arc models come from, and what do they assume?
Most current advice on arcs descends from the theatre and from screenwriting. Lajos Egri’s The Art of Dramatic Writing (1946) asks the writer to begin from a premise, which he defines as “a proposition antecedently supposed or proved; a basis of argument”, and gives Macbeth’s as “Ruthless ambition leads to its own destruction”. A play, for Egri, is the proof of its premise, carried out by characters whom conflict forces to change. Later manuals turned the proof into a sequence of named stages. John Truby’s The Anatomy of Story (2007) gives the hero a weakness and a need, a desire, an opponent and a self-revelation, the moment at which the hero sees the truth about themselves, and treats the story as a moral argument conducted through the hero’s and the opponent’s actions rather than through statement. K. M. Weiland’s Creating Character Arcs (2016) builds the arc on the lie the character believes at the start, which the story disproves, and names five arcs: a positive change arc, in which the character gives up the lie for a truth; a flat arc, in which the character already holds the truth and changes the world; and three negative arcs, disillusionment, fall and corruption.
The models are useful in one respect above all. They make the writer name the old criterion before drafting, as a weakness or a lie, so that the decisions that test it can be planned. Their cost is the assumption built into them, that change is the value a story serves and that it happens once, late, to a protagonist, in a sequence that can be scheduled. These are the conventions of twentieth- and twenty-first-century screen drama and of the commercial fiction that has learned from it. The picaresque, much modernist fiction, many long-running series heroes and a good deal of fiction outside the Anglo-American tradition do not arc in this sense, and the models will report them as failures. When they do, the models are describing their own assumption.
What is a theme, and how does a book argue one?
Egri’s premise and McKee’s controlling idea, which McKee describes as a value and the cause by which it is won or lost, both state a theme as a proposition. This chapter, following The Ending (Chapter 16), treats a theme instead as a question, and the difference is practical. A proposition written down before drafting tends to produce characters who illustrate it, and the opposition becomes a straw figure the book knocks down. A question written down before drafting can be answered by several characters in several ways, which is what Truby’s moral argument requires. The proposition is still useful, but as the answer a book reaches, found in revision by reading what the decisions argue, not as an instruction to the characters at the start.
What does the evidence say?
The evidence on how stories change readers is limited and indirect. Melanie Green and Timothy Brock (2000) ran four experiments on what they called transportation, the reader’s absorption into a narrative world. More transported readers held more story-consistent beliefs afterwards and judged the protagonists more favourably, and labelling a story as fact or as fiction made little difference to either. The finding supports a modest claim, that a story’s events can move a reader’s beliefs through absorption. It does not show that a stated theme persuades less than an enacted one. That is a conjecture consistent with it, since a statement addressed to the reader may interrupt the absorption the effect depends on, and the household reader’s Now I’m being told what to think is the kind of report the conjecture predicts.
See also: the full treatment, with the arc models set out in a table and the terms of each related to the criterion and the hinge.