The Same Spine That Makes Drama Falsifies History
The One Thing Roth Won't Hedge On
Eric Roth hedges about almost everything. Sixty years of practice, seven Oscar nominations, and through most of a long craft conversation he defers to taste, allows for other writers' approaches, softens his claims. Then he hits structure and the hedging stops cold: I don't care who you are, what you're doing. You can't break that.1
The "that" is three acts. State the problem. Complicate it. Show what the resolution costs. You can invert it, run it backwards, dilate one act and compress another — but the spine is the spine. It's the most rigid thing he says in the entire interview, and he's describing it as load-bearing for everything else: theme, character, dialogue, and scene work all hang off it. Strip the spine out and the rest collapses into unrelated tactics.2
Diarmaid MacCulloch, writing history at Oxford, holds the opposite discipline. His job is to refuse the spine, because real human events don't have act-two complications and act-three resolutions. They have ongoingness, partial closures, parallel timelines that never converge.3 Imposing the arc on them isn't craft. It's falsification.
Why This Isn't Just Two Genres Having Different Rules
The easy read is drama has one convention, history has another, everybody knows this. That undersells what's actually happening, because both writers are describing the same temptation and taking opposite disciplines against it.
The temptation is identical: material resists a clean shape, and the writer wants to give it one anyway. Roth's discipline is always give in, because dramatic form requires it. MacCulloch's is usually refuse, because historical truth requires it.4 Same pressure, opposite protocols — which means the choice isn't automatic and isn't determined by the material alone. It's determined by what the writer is claiming to have made.
And here's why the stakes are real rather than academic: the spine works. It genuinely does what Roth says it does. Narrative-transportation research measures how deeply an audience enters a story-world, and three-act structure is among the most reliable transportation-engineering tools available — stories that break it without compensating mechanisms measurably lose attention.5 Even more striking, the same load-bearing hinges show up in Sanskrit dramaturgy centuries before Aristotle, which suggests this isn't a Western screenwriting convention but something closer to how human attention sustains a story at all.6
So MacCulloch isn't refusing a mere convention. He's refusing the single most effective attention-holding device available to a prose writer, on purpose, because using it would make his book more compelling and less true.
The Two Failure Modes
Get the room wrong and you produce one of two specific, nameable failures.
Dramatized history — Roth's mode applied to MacCulloch's content. The historian who gives the material an act-one question, escalates it in act two, and lands a resolution. It reads beautifully. It's also making a claim about causation and closure that the evidence doesn't support, and the reader can't tell, because the shape is doing the persuading rather than the argument.
Arcless drama — MacCulloch's mode applied to Roth's content. The screenwriter who honors the messiness of real life and produces a film where act two is a montage of interesting scenes that don't escalate anything. Roth names exactly how this feels from the audience side: they can't tell whether events are building toward something or just wandering, tension dissipates, and the film "feels long" regardless of runtime, because no question is being held open.7
Both failures come from the same error — not knowing which claim you're making.
The Test That Actually Sorts It
The sorting question isn't is this fiction or nonfiction. Plenty of nonfiction legitimately uses the spine, and Roth's own field routinely adapts true events.
The sorting question is: what is the reader entitled to believe about causation when they finish?
If the answer is nothing — this is a made thing, and coherence is the contract, use the spine. Roth's Killers of the Flower Moon is a clean case, and it shows the spine doesn't even predetermine the ending. The act-one question is whether the law will protect the Osage. Act two escalates by revealing that the institutions built to deliver justice are themselves complicit. And act three refuses catharsis outright — Roth's closing note is that a jury will sentence a man more readily for kicking a dog than for killing an Indian.8 The spine didn't force a resolution. It made the absence of resolution land as a precise ending rather than a failure to finish.
If the answer is the reader should come away believing this is how events actually connected, then the spine is making an argument you haven't proven. Every hinge is a causal claim. Act two says this escalated because of that. Act three says this followed from those choices. In drama those are craft decisions. In history they're assertions about the world, and the shape smuggles them past the reader's scrutiny precisely because the shape is satisfying.
The Loose Thread Worth Pulling
Roth's own claim is empirically false as stated, and the vault says so directly: experimental cinema breaks three-act structure routinely, so does novel-of-consciousness fiction, so does much of the New Wave.9 His discipline is genre-specific — dramatic narrative aimed at general audiences — not universal physics.
Which opens the genuinely unresolved question. The vault names a third possibility neither writer occupies: a form using partial spine, something between full structure and full refusal, with its own equivalent of Roth's three sentences.10 The essay is the obvious candidate — Karlsson's version is described as sitting exactly in between, able to use three-act when the argument demands it without being bound by it.11 But nobody in this vault has specified what the partial-spine discipline actually is: what you commit to, what you refuse, and how you know when you've slid into either failure mode. That's a real gap in a form the vault otherwise covers in enormous depth.
Confidence
Roth's three-act claim, MacCulloch's anti-spine discipline, and the domain-split framing are all [VERIFIED] against the vault's own page, sourced to a named practitioner interview and tagged [PARAPHRASED] with line references at origin.13 The transportation research and the cross-cultural Sanskrit convergence are cited in the vault as supporting handshakes.56 The Killers of the Flower Moon three-act read is the vault's own analysis, not this report's.8 The sorting test proposed here — what the reader is entitled to believe about causation — is this report's synthesis; the vault names the two failure modes but doesn't supply a discriminator for choosing between disciplines in advance.
Threads: → The Same Spine That Makes Drama Falsifies History — Thread Map