Beau Lanier: Mateo, hey — I've been sitting with this weird thought all morning and I need to know if it makes sense to you or if I'm just, like, broken.
Beau Lanier: Every time someone says 'this decision caused that outcome' — every single time — they're already saying 'and if the decision had been different, the outcome would've been different too.' That's just... built in to saying anything caused anything. So why does the entire historical discipline treat the explicit version of that as unserious?
Mateo Rivas: No, that's not broken — that is actually the epistemological knot we're untying today. Because counterfactual history, the formal 'what if?' practice, gets called speculation. But every causal claim a historian makes already embeds a counterfactual. The crime is apparently saying it out loud.
Beau Lanier: Hold on — so the problem is transparency, not the reasoning itself?
Mateo Rivas: That's the tension, yeah. And it crystallized in a very public way — 2014, the WWI centenary, Niall Ferguson on television arguing that Britain's non-entry into the war would have been manageable, Max Hastings firing back that Germany would have rolled out something resembling a proto-Nazi dictatorship across Europe. Two serious historians doing exactly what historians do in their footnotes, except visibly, and the discipline largely recoiled.
Beau Lanier: That's a huge gap — like, Ferguson's saying 'it would've been fine,' and Hastings is saying 'it would've been fascism.' Those are not small disagreements.
Mateo Rivas: Right — but here's what that gap actually shows you. Both of them had to commit to a causal mechanism, a verifiable chain of decisions and structures, to get to their conclusion. Ferguson's reading WWI as a contingent disaster Britain accelerated. Hastings is reading German military ambition as structural, not contingent. The counterfactual forced the assumption into the open.
Beau Lanier: It's like — okay, sorta like two friends arguing about why their buddy's marriage didn't work. 'If he'd stayed in that city, they'd have made it.' 'No, it was already falling apart, the city was just when it showed.' You can't prove either one. But the argument surfaces what each of them actually thinks the real cause was.
Mateo Rivas: Pues, that's the whole thing. The 'what if' is the shadow cast by every causal claim — you can't have one without the other. So the real question driving this whole episode isn't 'is counterfactual history legitimate.' It's: why is the discipline so committed to keeping that reasoning invisible? Are they protecting knowledge, or are they protecting a very specific permission structure about what knowledge is allowed to look like?
Beau Lanier: And — I mean, those feel like very different things with very different answers.
Mateo Rivas: They are. And the evidence that they might be getting it backwards is genuinely unsettling once you see it.
Beau Lanier: Backwards — that's the word that keeps snagging me. Like, okay, imagine a grad student writing a dissertation sentence. Something like, 'Bismarck's decision to isolate France diplomatically created the conditions that made WWI possible.' That's a sentence in a respectable thesis. No one blinks. But — wait, is that student doing counterfactual history or not?
Mateo Rivas: Pues, technically? Yes. The sentence only works if you believe that without the isolation, those conditions don't exist.
Beau Lanier: Which is — okay, so there's this paper, Matthew Rellihan, 2025, Journal of the Philosophy of History. And what he actually does is build a threefold taxonomy. Like, not 'counterfactuals: good or bad' — he cuts it into three grades. The first grade is what that dissertation sentence is already doing. He calls them unproblematic counterfactuals, the ones that are just embedded inside ordinary causal claims. And those are already in every respectable history book ever written.
Mateo Rivas: Hold on — every history book?
Beau Lanier: Every. One. Because the causal mechanism — y'know, the chain of decisions and structures historians reconstruct from evidence — that chain only means something if you accept that removing any link changes the outcome. That's not speculation, that's just... what causation is.
Mateo Rivas: Rellihan's second and third grades — moderately speculative, then hopelessly speculative — those are where 'what if Napoleon won at Waterloo' lives?
Beau Lanier: Right, and the problem he identifies is that there's no bright line. Like, it's degree, not kind. The Bismarck dissertation sentence and the 'what if Napoleon won' question are on the same spectrum — you can't draw a clean wall between them. The discipline acts like there's a wall.
Mateo Rivas: That's — bueno, that's structural inconsistency, not just sloppiness. They're actively using the logic they're publicly disowning.
Beau Lanier: Yeah, and I mean — the inconsistency isn't accidental, right? It's baked in. But here's where I want to slow down, because I don't think the firewall is purely defensive nonsense. There's actually a reason someone built it.
Mateo Rivas: What reason could justify pretending the logic isn't already there?
Beau Lanier: The evidence problem, y'know? Like — that Bismarck sentence is anchored. There are state papers, diplomatic cables, actual archives of what Bismarck actually did. Once you start sliding up Rellihan's spectrum toward 'hopelessly speculative,' you leave the archive behind entirely. So the firewall isn't wrong about the risk. It's just — I mean, it's drawing the line in the wrong place. It's banning the transparency instead of banning the recklessness.
Mateo Rivas: So the discipline built a rule that solves the right problem but catches the wrong thing. Protects evidence-based reconstruction by treating the visible form of the reasoning as the threat — not the actual departure from evidence.
Beau Lanier: But — wait, so, catching the wrong thing is still catching something real, right? Like, I don't want to just walk past the fact that the firewall exists for an actual reason.
Mateo Rivas: No, claro — and this is where I want to make the strongest case for the other side, because it deserves it. Picture a historian, actual historian, sitting at a reading room table at the Archives nationales in Paris. Box of Napoleonic-era correspondence in front of her. She pulls a letter, reads a claim, writes an interpretation. The next letter in the same box can prove her wrong. That's what falsifiable means in that room — not a philosophy seminar concept, a Tuesday morning in an archive.
Beau Lanier: Okay, that's — that's actually a really precise image.
Mateo Rivas: And the Stanford Encyclopedia of Philosophy grounds the entire epistemology of history in exactly that: only claims checkable against surviving traces qualify as historical knowledge. Marek Tamm says the same thing from a pragmatist angle — credibility is anchored to that evidentiary discipline. Both of them are pointing at the same Tuesday morning. Pull the letter, check the claim.
Beau Lanier: So Ferguson — is he just doing elaborate fiction, then? Like, that's what the critics were saying?
Mateo Rivas: That's exactly what the 2014 Guardian piece argued, yeah. That the whole virtual history exercise — his term, virtual history — distracts from explaining what actually happened through evidence. And bueno, I think that critique is honest. Because no box in any archive contains the correspondence from the timeline where Napoleon won at Waterloo. That paper does not exist. The revisability that makes the Archives nationales historian's work knowledge — that mechanism is just gone.
Beau Lanier: No archive of things that didn't happen.
Mateo Rivas: None. So the firewall isn't irrational — it's tracking something real about where knowledge comes from. The problem is it gets wielded to police transparency, not to police the actual departure from evidence.
Beau Lanier: But — okay, here's what's nagging me. That historian in Paris, pulling letters out of a box. She's reading those letters knowing how everything turned out. Napoleon lost. She knows that going in. So is she — I mean, is she actually reading the evidence clean, or is she reading it through the outcome?
Mateo Rivas: Pues... that is the complication that breaks the clean version of this argument.
Beau Lanier: Like — the falsifiability test says the next letter in the box can prove her wrong. But what if the way she's reading every letter in the box is already bent by knowing the ending? That's not a problem the archive can fix from inside the archive.
Mateo Rivas: And that is exactly where this gets genuinely uncomfortable — because there's a 2019 Harvard Kennedy School paper, Mukharji and Zeckhauser, on what they call explanation bias. The measurable tendency to treat past outcomes as more inevitable than they were. And what their finding suggests about counterfactual thinking — that's the part that should make everyone who just defended the firewall sit down for a second.
Beau Lanier: So the tool that can't produce knowledge might be — wait, we're going to sit with that one, because that's not a small thing.
Mateo Rivas: Mukharji and Zeckhauser documented it in 2019 — they called it explanation bias, and the mechanism is specific: once you know the outcome, you retroactively inflate how probable that outcome was all along. Not just a little. Measurably. So the historian pulling letters in Paris isn't just framing the evidence — she's actually misreading the prior probabilities embedded in the document itself.
Beau Lanier: Wait — misreading the actual numbers in the document?
Mateo Rivas: The judgments, yeah. Okay, think about a medical examiner reviewing a patient's chart after the patient died. Every decision in that chart — the dosage, the missed symptom, the delayed referral — now reads as obviously leading to death. The narrative has a gravitational pull toward the ending you already know. But a colleague who comes in blind, doesn't know the outcome, reads the same chart and goes: there were three plausible branch points here. That second person is doing implicit counterfactual thinking. And the research says they're seeing the evidence more accurately.
Beau Lanier: The colleague who doesn't know the ending is — I mean, that's the one we'd trust in a malpractice case. That's the one whose read we'd actually want.
Mateo Rivas: Exactly. And the paradox Mukharji and Zeckhauser are sitting with is: the tool that cannot produce historical knowledge, because it's unfalsifiable, no archive behind it — that tool might be the only guard against misreading the archive we do have.
Beau Lanier: That's — I don't know how to sit with that, y'know? Like, the thing you banned from the room is the thing keeping the room honest.
Mateo Rivas: And it connects back to historical contingency — the idea that past outcomes weren't inevitable. Because if you can't hold onto contingency while you're reading the evidence, explanation bias fills that space. The inevitable-looking story writes itself right over the actual record.
Beau Lanier: Okay but — wait, here's what I'm sorta stuck on. Gavriel Rosenfeld's work. His read on popular alternate history is that those narratives — the 'what if the South won,' the nightmare scenarios — they're actually about the present, not the past. Like, nightmare alternate histories vindicate the world we have. Fantasy ones critique it.
Mateo Rivas: Presentism. The genre is more about now than then.
Beau Lanier: Right — so if Rosenfeld is right, then the 'hopelessly speculative' end of Rellihan's taxonomy isn't broken because it leaves the archive. It's serving a completely different function. It's doing cultural and psychological work, not epistemological work. And neither dismissing it entirely nor calling it history gets at that.
Mateo Rivas: Bueno — so there's the split. Counterfactual thinking close to the evidence, the kind that corrects explanation bias, that's epistemically necessary. The 'what if' that drifts into serving present anxieties — that's a different thing with real value, just not historical knowledge.
Beau Lanier: Which means — I mean, the discipline's mistake isn't drawing the line. It's drawing it at transparency instead of drawing it at, like, evidential distance. Ferguson and Hastings on television, both of them tethered to actual causal mechanisms — that's different from wish-fulfillment alternate history. Those aren't the same activity.
Mateo Rivas: And the Mukharji and Zeckhauser finding makes that distinction load-bearing — because if explanation bias is real and measurable, then the cost of banning all counterfactual thinking isn't just methodological tidiness. You're leaving the archive exposed to the very distortion you built the firewall to prevent.
Beau Lanier: Y'know what that means for the dissertation student, though — the one writing the Bismarck sentence. She's not doing anything wrong by making it explicit. She's just... being honest about the reasoning that's already in there.
Mateo Rivas: That's the whole landing, I think. Rellihan's 2025 taxonomy — what it actually opens up is a reformed methodology where you keep the unproblematic counterfactuals, the ones embedded in every causal claim, and you constrain them to the evidence. You're not becoming Ferguson writing virtual history. You're showing your reader exactly which link in the causal chain has to hold for the whole argument to stand.
Beau Lanier: So it's not 'counterfactuals yes or no' — it was always just 'how visible are you willing to make the thing you're already doing.'
Mateo Rivas: And the discipline chose invisible. Which is — bueno, that's not an epistemological wall. That's a choice about risk tolerance and credibility optics. Because showing the counterfactual reasoning makes the claim more falsifiable, not less. Your reader can now see what evidence would have to be different to break it.
Beau Lanier: Wait — more falsifiable. That's the part that kind of stops me cold, man. The thing that looks like speculation is actually the more rigorous move.
Mateo Rivas: That's where I landed too. And it takes you all the way back to what you said at the top — every causal claim already has a counterfactual built in. Turns out the crime was always saying it out loud. Not the reasoning. Just the transparency.
Beau Lanier: Mm. We started with that weird feeling that I couldn't shake — and I think what I found is that the feeling was right, it was just pointing at something more specific than I expected.
Mateo Rivas: Pues — I'm glad you said it out loud this morning. Good conversation.
