Onpode
Cover art for When could history have gone differently—and when was the outcome overdetermined

When could history have gone differently—and when was the outcome overdetermined

November 15, 2026 · 15 min

Clara Bennett & Finn Brooks

Finn, hey — I want to start with something uncomfortable: I think we've been teaching causation wrong. Oh that's — wait, what do you mean wrong, like factually wrong or framing wrong? Framing wrong. We hand people a single trigger and call it a cause. And the case that exposes that completely is the assassination…

The structure-agency debate is one of the most foundational and enduring tensions in historical interpretation.

0:0015:26
Get the next episode on History →

Follow it free — new episodes land in your feed.

Or make your own — any topic, in minutes

More Onpode episodes on History →

About this episode

The structure-agency debate is one of the most foundational and enduring tensions in historical interpretation.

Grounded in 11 sources
Contingency and History | Philosophy of Science | Cambridge Core ↗ · cambridge.org
Narrative and Its Non-Events: Counterfactual Plotting in the Victorian ... ↗ · dash.harvard.edu
Causal Explanation in History: An Analysis of Research Practices ↗ · doi.org
On the “Duel” Nature of History: Revisiting Contingency versus Determinism ↗ · doi.org
Contingency and historical inevitability ↗ · doi.org
A Historical Review of Counseling Theory Development in Relation to Definitions of Free Will and Determinism ↗ · doi.org
Historians and the French Revolution ↗ · doi.org
Free Will, Determinism and the “Problem” of Structure and Agency in the Social Sciences ↗ · journals.sagepub.com
Narratives of contingency as historical evidence for philosophical arguments of contingency: pathway decisions in the early development of molecular genetics | History and Philosophy of the Life Scien ↗ · link.springer.com
The Size of History: Coincidence, Counterfactuality and Questions of Scale in History | Springer Nature Link ↗ · link.springer.com
Contingency from Braudel to Blumenberg ↗ · rofl.stanford.edu
Read transcript

Clara Bennett: Finn, hey — I want to start with something uncomfortable: I think we've been teaching causation wrong.

Finn Brooks: Oh that's — wait, what do you mean wrong, like factually wrong or framing wrong?

Clara Bennett: Framing wrong. We hand people a single trigger and call it a cause. And the case that exposes that completely is the assassination of Archduke Franz Ferdinand — June 28th, 1914, Sarajevo.

Finn Brooks: Okay this is exactly what I want to dig into — because I've been kind of obsessed with the actual mechanics of what happened that day. Like, Gavrilo Princip's first shot missed. The whole thing should have been over. And then the driver takes a wrong turn — just, logistically screws up the escape route — and Princip ends up standing directly in front of Franz Ferdinand again. Second shot lands. World War One starts. That's not — I mean, that's not a cause, that's barely even a plan.

Clara Bennett: Right — but here's the structural counter, and it's genuinely strong: the alliance systems, the arms race, the imperial competition — historians argue a major European war was already highly probable. Princip found an opening; the opening existed whether or not he found it.

Finn Brooks: Which is — okay I love that argument, BUT — it doesn't explain the form the war took, or the timing, or which specific alliances activated when. Like, does the structure tell you *that* or just that something big was coming?

Clara Bennett: That is the question. And that gap — between 'the trigger' and 'the cause' — is what today is actually about.

Finn Brooks: Because you can hold both things at once — the wrong turn was contingent, it really could have gone the other way, and war was structurally probable. Those aren't — wait, actually those aren't contradictions, are they.

Clara Bennett: They're not. That's exactly the tension worth not resolving too quickly.

Finn Brooks: Okay but — not contradictions, sure, but then which one do we actually *call* the cause? Like, in practice, when someone asks why World War One happened, do you say the alliance systems or do you say Gavrilo Princip?

Clara Bennett: That's the wrong question, and I want to show you why with something concrete. Think about an icy road. The ice doesn't make a crash certain — plenty of people drive on ice without crashing. But it makes a crash very likely. Then one driver swerves, one specific moment, and there's the crash. Now: what caused the accident?

Finn Brooks: I mean — both? The ice and the swerve?

Clara Bennett: Both. And that's the whole debate in one breath. The ice is the distal condition — it's structural, it built up over time, it was doing most of the work before anyone got in the car. The swerve is the proximate trigger — specific, real, happened at a moment. Conflating them is a classic historical error. Princip's shot is the swerve. The alliance systems, the arms race, imperial rivalry — that's the ice.

Finn Brooks: Oh. Okay that — wait, that actually lands. So historians who fight about this are really just... pointing at different parts of the same crash.

Clara Bennett: Structurally, yes. Fernand Braudel — and this is his actual move, not just the phrase people cite — he separates it into layers. Deep structural time, centuries long. Medium cycles, decades. And then the event layer, what he calls the surface. Princip firing that second shot on June 28th, 1914? Surface. The ice under everything? That's been building for a century.

Finn Brooks: So does Braudel think the shot basically doesn't matter?

Clara Bennett: He'd say it matters — but only at its own level. The spark matters. It just doesn't explain why the room was full of gas.

Finn Brooks: Right right — but the gas doesn't tell you *when* it goes up, or how, or which specific alliance activates which ultimatum in which order. That detail — that's not in the ice.

Clara Bennett: Exactly. And that's what contingency theorists hold onto — the specificity. Alter one variable on that Sarajevo street, the driver doesn't take the wrong turn, and maybe the form the war takes, the timing, the particular chain of mobilizations — those change. Not 'no war ever,' maybe, but *this* war, the one we got, is not inevitable.

Finn Brooks: So does it matter which one we call the cause? Like — practically. If you're a historian writing the textbook, does getting this wrong actually do damage?

Clara Bennett: It does. If you teach only the proximate trigger, students think history turns on drama and accident. If you teach only structure, individuals disappear and nothing specific ever seems preventable. The real work is being precise about which level you're claiming causation at. That's the question historians keep skipping.

Finn Brooks: And that precision thing — which level are you claiming — that's actually what Braudel builds the whole system around, right? Because it's not just 'zoom out and structure wins.' It's more like... the timescale you pick determines what counts as a cause at all.

Clara Bennett: That's exactly it. And the counterintuitive part — the part people miss — is that zooming out doesn't eliminate contingency. It changes the question you're asking.

Finn Brooks: Wait — say that again.

Clara Bennett: Let's say you're a supply-chain analyst, 2008. Your company just collapsed. Zoomed to that specific week — your CEO made a panicked call, pulled the credit line on a Tuesday morning, and that decision cascaded. That call was decisive. It happened. It mattered. Zoomed out to the decade, the financial architecture — the leveraged instruments, the exposure — made some collapse highly probable. Both are true. Neither cancels the other.

Finn Brooks: Okay but — hang on — if the decade-scale says collapse was probable, doesn't that make the CEO's Tuesday call kind of... noise? Like, Braudel's 'surface foam'?

Clara Bennett: Only if you're asking a decade-scale question. If you're asking why this company collapsed before that one — why Tuesday and not Thursday, why this credit line and not another — the CEO's call is not foam. It's the whole answer. Braudel's three levels aren't competing explanations. They're different magnifications of the same reality.

Finn Brooks: Different magnifications — not different realities. Okay, that's actually a really clean way to put it.

Clara Bennett: Now, there's a philosopher who pushes back on Braudel's whole framing — not on the facts, same facts — Hans Blumenberg. Where Braudel embeds events inside structure, Blumenberg says the openness of individual moments is irreducible. You can't dissolve it upward into the century-scale. The moment the CEO picks up the phone, that's genuinely open — not just analytically open, ontologically open.

Finn Brooks: So Blumenberg is basically saying Braudel is doing something sneaky — like, packaging contingency inside structure until it disappears?

Clara Bennett: Not sneaky — different framework. Braudel isn't claiming the CEO had no choice. He's asking which analytical level illuminates the most. Blumenberg says that move costs you something real. The two of them aren't arguing over history — they're arguing over what the act of historical explanation is for.

Finn Brooks: And then there's Nigel Pleasants — wait, the argument is that this whole fight is partly a category error? Like, the metaphysical question and the empirical question keep getting run together?

Clara Bennett: Pleasants says if you separate them — if you stop asking 'is history determined in principle' and start asking 'which factors did the most causal work in this specific case' — a lot of the deadlock just dissolves. Historians keep fighting because they're answering an empirical question while reaching for a metaphysical vocabulary. And actually — the tool that makes that empirical question feel answerable is the thing that can go very badly wrong. We'll get into counterfactual reasoning next, what it actually proves versus what people think it proves.

Finn Brooks: Oh — yeah, because 'what if the gun misfired' feels like evidence but I'm not sure it actually is.

Clara Bennett: And that instinct — that 'what if the gun misfired' feels like evidence — is exactly where counterfactual reasoning gets dangerous. Because it's the most powerful tool historians have for isolating causal weight, and the most easily abused.

Finn Brooks: Okay but wait — why is it dangerous? Like, isn't asking 'what if X hadn't happened' just... how you figure out if X mattered?

Clara Bennett: It is — but only if you hold everything else constant. That's the legitimate version. You change one variable, the most proximate one, and you stay close to what was actually possible. Imagining Princip's gun jamming on June 28th — that's minimal, it's historically proximate, it's legitimate.

Finn Brooks: Right, because a gun can misfire. That's not a fantasy, that's just — mechanical probability.

Clara Bennett: Exactly. Now imagine instead: what if there had been no European imperial rivalry at all? That's not isolating a contingent trigger anymore — you've changed the structural conditions entirely. You've run a different civilization. That counterfactual tells you nothing about Princip.

Finn Brooks: So the illegitimate version just — treats any imaginable alternative as equally plausible. Like, 'what if nationalism never existed' is not the same kind of question as 'what if the driver hadn't taken the wrong turn.'

Clara Bennett: And here's where my actual skepticism lands — because even the legitimate version, the gun-misfires version, doesn't prove what people think it proves. If you remove Princip and war still happens six months later, that tells you something about how tightly wound the structural conditions were. It does not tell you Princip was causally irrelevant.

Finn Brooks: Wait — say that again, because that feels like a real crack in the method.

Clara Bennett: He may have shaped the war's form, its speed, which specific alliance activated which ultimatum first. We collapse 'caused the outcome' with 'was essential to the outcome' — and those aren't the same thing. Counterfactual reasoning isolates a variable by assumption, not by evidence.

Finn Brooks: Okay that is — no, I mean, that actually breaks the whole Marxist-structuralist move too, doesn't it? Because structuralist historians will say individual leaders only alter timing or form, not fundamental direction — but that's basically doing a counterfactual where they've already decided the structural conditions dominate. They've smuggled the conclusion in.

Clara Bennett: Mm. The structuralist counterfactual collapses exactly there — because class relations and economic change explain that *a* revolution happens, not that *this* revolution, in *this* form, with *these* specific casualties and outcomes, was mandated.

Finn Brooks: And there's actually a case from science — this is the part that genuinely got me — philosophers reconstructed the fork-moment decisions scientists made during the development of molecular genetics. Actual archival sources, actual choice points. And the argument is: the specific pathway taken was not structurally mandated. There were live alternatives. Real ones, not fantasy ones. That's not a historian speculating — that's the legitimate version of the method, applied rigorously.

Clara Bennett: And that's what 'narratives of contingency' actually means as a concept — it's dual. The scientists themselves, in real time, constructed prospective narratives about which path to take. And then historians reconstructed those same fork moments from the archive. Two different uses of the same framework. That's the method working as it should.

Finn Brooks: So the tool isn't broken — it just has really strict operating conditions. Stay proximate, hold structure constant, and — wait, actually, don't confuse 'this trigger mattered' with 'this trigger was the only thing that mattered.' Those are three different claims and we keep running them together.

Clara Bennett: And that's actually where I want to land — because David Christian has this framework, the S-curve, where population growth, technological diffusion, state formation — they all follow the same shape. Slow start, rapid acceleration, plateau. That's the structure. Real, measurable, not invented. But the S-curve doesn't specify which path gets taken inside it.

Finn Brooks: So the shape is determined but the... route through it isn't.

Clara Bennett: Exactly. Contingency fills the gap the structure leaves open. That's not a compromise between the two camps — it's just an accurate description of how the thing actually works.

Finn Brooks: Which means — and I think this is the click for me — asking 'did Gavrilo Princip's bullet matter' a hundred years later isn't actually an answerable question until you say which timescale you're asking from. Braudel's century-scale? Maybe war was coming regardless. The week of June 28th, 1914? That bullet is everything.

Clara Bennett: And most historians now — they're not at either extreme. The serious work is in the interaction. Where does the structure constrain, and where does the contingent moment actually redirect?

Finn Brooks: Right — but which one you emphasize, that's not purely an evidence call, is it. It's also... who's asking, and why.

Clara Bennett: The causal question, the audience, the timescale — those shape the emphasis before the evidence even enters. That's not a flaw in the method. It's just being honest about what explanation is doing.

Finn Brooks: We started with 'we've been teaching causation wrong' — a single trigger, hand it to students, call it a cause. And I think where we ended up is... the trigger isn't wrong to teach. The wrong part is pretending it's the whole story without saying which level of the story you're actually in.

Clara Bennett: That's it. Princip's shot is real. Braudel's ice is real. The honest move is just saying which one you're pointing at and why. That feels like enough.