I’ve become fascinated by a specific kind of counterfactual argument that is often made, but is constraining and damaging in deliberation. Let’s call it the “necessary counterfactual.” It’s structurally unstable, but people tend not to see that, and often find it compelling. And its weaknesses are hard to explain.
Just to be clear: I’m not arguing the counterfactuals are always bad—they’re inescapable in policy deliberations, whether personal or “political.” Any causal argument implies a counterfactual (if I say that you feel sick because you ate a party-size bag of knock-off chocolate-covered fried bananas, I’m implying that you wouldn’t feel sick if you hadn’t).
The problems arise when the connections are asserted and assumed to be so necessarily connected that all dissent or disagreement is delegitimated.
For instance: When confronted with an expansionist and aggressive leader (Hitler) decision makers (the French and British) chose appeasement. Appeasement didn’t work; therefore, it is obvious that maximum aggression would have worked. We are now in a similar situation (Korea, Vietnam, Iraq), so it is obvious that we must do what would have worked then, and engage in maximum aggression.
Here’s the structure of the logic:
In [Situation Z], policymakers adopted [Policy A], and [Policy A] went badly; therefore, the correct policy must have been [Not-A]. [Policy B] is not-A; therefore, it is obvious that [Policy B] would have worked. We are now in a situation similar to [Situation Z]; therefore, it is obvious that we must adopt [Policy B], since that would have worked the prior time.
The “logic” is something like this:
A –>B; therefore ~A –>~B.
(A led to B, therefore not-A leads to not-B.) Appeasement led to failure; therefore, not-appeasement would have led to success.
What makes the “logic” of this kind of argument so troubling is the degree to which it assumes necessary consequences and necessary implications—the way that it shuts down reasonable disagreement.
This kind of “logic” ignores several important points:
1) That A didn’t work doesn’t necessarily mean that this not-A would have worked.
That sentence has a lot of negatives (which is probably why it’s hard for people to grasp). But my point is that this way of thinking assumes two false binaries that are falsely assumed to be necessarily connected.
There are plenty of events/policies other than maximum aggression that might have prevented Hitler’s coming to power, starting WWII, engaging in the multiple genocides: his getting into art school, von Hindenburg and von Papen refusing to work with him, the Reichstag growing a spine and refusing to pass the Enabling Act, anti-Nazi parties making a coalition, one of the assassination attempts working, a treaty that wasn’t as punitive as the Versailles Treaty, the US joining the League of Nations, economic policies that would have meant the Great Depression didn’t happen, Germany refusing to back Austria’s play in 1914, a different response on the part of the French in May of 1940, Napoleon III not taking the bait and declaring war in 1870…
2) Evidence isn’t proof.
Counterfactuals are generally supported by telling a story about what would have happened had the other course been taken. That maximum aggression would have worked with Hitler is argued through the telling of a compelling narrative. But that it’s compelling doesn’t mean it was likely, let alone guaranteed.
One narrative is that a military response to Hitler’s remilitarizing the Rhineland would have led to a military coup that would have removed Hitler from power. But the only evidence that a coup was even considered are things said in self-serving and jaw-droppingly dishonest post-war memoirs on the part of Nazis military leaders, who lied about everything else (such as the Wehrmacht’s participation in genocide), so why suddenly decide they’re being honest about that coup plot? Particularly since there is no other supporting evidence. A military response, which Hitler thought was likely, wouldn’t have stopped him from regaining that territory (they were going to get it anyway in a few years), and it wouldn’t have persuaded him to change his mind about his long-term plans. What might have made a difference would have been a more successful response to the Nazi invasion of France, and there are counterfactuals about what would have been better responses, but even a compelling counterfactual is still a counterfactual.
3) Assuming best case counterfactual scenarios isn’t even evidence.
Richard Ned Lebow makes this point quite neatly in Forbidden Fruit. Counterfactual narratives often make one major change, and then assume that there are necessary consequences of that change, all of which are the consequences most favorable to the argument. So, a military response to Hitler’s actions in the Rhineland would have successfully deterred him, there would have been a coup, it would have been successful (a very weak link in the chain of events), the Nazi military leaders would have put in place a non-aggressive, non-expansionist, and pacific government that would not have tried to refight the Great War.
There are other stories one might tell that are also plausible: it might not have deterred him from invading Czechoslovakia; it might have increased anti-war sentiment in France and Britain; there probably wouldn’t have been a coup; it was very unlikely that a coup would have been successful; Nazis would still have been Nazis.
4) Policies have to be politically available.
Many scholars argue that going to war over the Rhineland was not politically possible for France or Germany. It would have seemed to be a repeat of the Great War, which was (in the 20s and 30s) often narrated as a massive overreaction to a minor incident with no real importance.
5) That A didn’t work doesn’t necessarily mean that there is a not-A that would have worked, one which must be our current choice.
I’m not making some argument about history being determined, or non-contingent; I’m making a point about argument, but I’ll get to that. I’m saying that history is contingent and options are contextual. For instance, there are counterfactuals that I’m persuaded would have prevented WWII, at least in terms of the form it took. I’ll mention two: if Hitler had gotten into art school; if the Franco-Prussian War hadn’t happened.
If we follow the logic of the necessary counterfactual, then the implication of the first is that we should let everyone into art school. Or maybe just all Nazis? Or something.
What are the policy implications of the second counterfactual? Don’t go to war against Bismarck? That’s probably pretty good advice, but unlikely to be useful, and certainly not something that might have helped people trying to figure out what to do about Hitler. Or Vietnam.
So, counterfactuals about the past, even very plausible ones, don’t necessarily determine our current policy choices. We’re not in the realm of obviously correct decisions, but in a world of deliberation.
The problem with necessary counterfactuals is not that they’re counterfactual, but that they’re presented as necessary at every step—this is necessarily what would have happened, and that narrative makes it clear what is necessary for us to do. My point is about how they shut down potentially useful disagreement. My crank theory is that’s exactly why people are drawn to them—they don’t actually reduce uncertainty, but they make us feel less uncertain about the past, present, and future. The narrative about obvious choices is comfortable, since it ends the need to acknowledge uncertainty and risk.
And, obviously, I’m not just talking about Vietnam. Or Hitler. There are a lot of discussions right now about what went wrong such that we have a political party that simply refuses to hold its members accountable, that is openly pursuing competitive authoritarianism and abandoning democracy, and the question is: what should opponents of authoritarianism do now? And that question means thinking about how we got here—it involves counterfactuals. And it should. But what won’t help is swearing fealty to a necessary counterfactual. I mean, it might help people feel less uncertain, but it won’t help figure out our various options.
