The Uses of Good Counterfactual History
More than a "parlor game" and less dead than you think -- if you know where and how to look
Earlier this week at Slow Boring, Matt Yglesias published a stimulating essay on the value of counterfactuals, with a modest plea for their utility as in understanding “causal claims” — how and why things happen — that begins with an exploration of the alternate history genre.
Counterfactual history, which explores how history might have unfolded differently if key facts were changed, clearly makes for thought-provoking, entertaining, and sometimes even good fiction. I think in particular of two favorites: Laurent Binet’s Civilizations (2021), in which the Incans come to the “Old World” and upend European politics and culture, and Richard Harris’ Fatherland (1992), a detective novel set in a mostly victorious but brutal, stagnant, and decaying Third Reich in the 1960s.1
But, as Yglesias notes, counterfactual history is also slightly disreputable among professional historians who often describe it as fundamentally unserious. This charge is true enough, but I would qualify the conclusion — and offer some initial thoughts on where to find interesting work in the field.
The Charge: The Problem of “What If”
It is easy enough to see why a certain kind of historian would object to counterfactuals — and to the people who rise to prominence through them. History, after all, is a discipline which gets a lot of love but very little respect from most laymen. The professional historian constantly runs into people who profess to be deeply interested in history, but turn out to be Kennedy assassination conspiracy theorists or readers of popular histories of World War II with idiosyncratic (if not outright incorrect) views of Pearl Harbor, the Wehrmacht, the Manhattan Project, or the decision to drop the atomic bombs on Hiroshima and Nagasaki.
Counterfactual history can come across as “a parlour game with the might have beens of history” as E.H. Carr, the great historian of international relations, famously put it. As practiced by many prominent writers, the counterfactual can simply become a vehicle for fantasy that is very far from the kind of meticulous analytical and archival work that most historians do. Hence why David Hackett Fischer, famous for his painstaking work on the British settlement of North America and its influence on folkways, condemned it Historians’ Fallacies (1970) — his takedown of peers in the profession for various sins of omission and commission.
Counterfactual history also gives license for unfalsifiable speculation that usually says more about the writer’s political predilections than the evidence at hand. One can see why Richard Evans might condemn it outright in his book Altered Pasts: Counterfactuals in History (2014) as the work of those treading “on thin evidential ice.” (Considering that Evans served as an expert witness against amateur historian, serial fabricator, and Holocaust denier David Irving in the libel trial of scholar Deborah Lipstadt, one can see why he might be particularly vituperative when sensational writers get creative with the past.)2
Once one steps away from actual evidence, it becomes very easy to let one’s imagination run wild — with speculation and the mere repetition of one’s conclusions hardening into conventional wisdom that no amount of debunking can counter. Just witness the spectacle Bill O’Reilly and Martin Dugard’s best-selling Killing Reagan which is littered with inaccuracies and distortions that make it basically history as fantasy, albeit of a deeply damaging kind.
When the facts become excessively stylized and narratives become untethered from what happened, it gets very easy to draw the wrong conclusions — as Yglesias arguably did in recent speculation about how the Republican Party and history would have changed had Teddy Roosevelt been elected in 1912 on Twitter, with the questionable assertion that this change would have kept African Americans in the Republican tent rather than becoming part of the New Deal Coalition, a contention that I pushed back on a bit.
This example illustrates the tension between useful speculation and retrospective wish-casting. There is a reason that alternate history writers focus on prominent singular personalities and major events (the American Civil War and World War II being the most prominent), but it’s easy to lose the thick texture of the times. A closer study of TR’s political campaigns, approach to domestic reform, and overall policy priorities does not necessarily suggest that he would have been more progressive (Yglesias’ argument, essentially) in practice than Republican contemporaries on civil rights or even on economics.
Consider the contrast between TR and Charles Evans Hughes, Wilson’s Republican opponent in 1916, drafted pro-civil rights opinions as an associate justice on the Supreme Court and campaigned against lynching. TR, a less prejudiced man than many in the early twentieth century, opted to run a “lily white” campaign in the South, seating all-white delegations at the Progressive Party convention rather than the biracial delegations that also demanded to be seated. He arguably hurt himself in the South by publicly dining with African Americans in the North during the campaign, which suggests his egalitarianism, but TR was willing to make tactical accommodations and could also make tactical mistakes — even though he hoped to change the electoral map.
This leads to an additional question. Would TR, a man most compelled by leadership on the world stage and acutely aware of the need to carefully manage public opinion when it came to action overseas, have spent political capital as president from 1912 to 1916 on civil rights or even extensive domestic legislation, especially after 1914? Given TR’s enthusiasm out of office for military preparedness and entry into the Great War, it is not necessarily clear that he (or someone like Hughes, had he been elected in 1916, for that matter) would have cashed in chips to make the major changes Yglesias suggests that would be needed to prevent racial realignment under FDR’s New Deal coalition in the context of the Great Depression.3 In short, the counterfactual is not quite as compelling as it may have initially seemed.
The Counterfactual Defended: The Value of “If, Then” When Bounded
The point here isn’t simply to poke holes in speculation on Twitter, but rather to suggest that counterfactuals depend upon a unique combination of knowledge about an era and an openness to the possibility of change. Even a counterfactual debunked can have value, because it forces us to encounter facts, debate interpretations, and articulate causal mechanisms with greater clarity. And good counterfactuals can reveal the relationship between contingency (the possibility of the unforeseen due to chance and human choice) and structural factors that constrain the course of events as few other methods can.
As Yglesias puts it, we’re all doing counterfactual history all the time in the sense that “every causal claim is, implicitly, a counterfactual claim.” Cass Sunstein made a similar argument to even greater effect in 2014 in a worthwhile essay in The New Republic, later expanded in an academic journal. Counterfactuals are inevitable, whether in the writing of history or in our personal lives, because when we say that something is “necessary” for a given event we imply that it wouldn’t have happened otherwise, thus creating a “parallel world.”
So, more succinctly: counterfactuals can help us discern what truly matters in a given situation. What was sufficient for something to happen versus merely necessary? What was overdetermined and would have been the same regardless? What were the actual “hinge moments” where history might have gone in a different direction?
Fortunately, once you look outside of traditional academic departments and past the requisite throat-clearing that inevitably comes up when historians talk about concepts such as presentism, reasoning via historical analogy, the application of history to policy, or teleology, the value of counterfactual history is obvious. And there’s more of it going on than is sometimes apparent — especially in the fields of diplomatic, military, and political history.
Consider a few turning points that I’ve come across in recent articles and conversations:
If Gavrilo Princip had decided to pack it in on June 28, 1914, after the first assassination attempt on the Archduke Ferdinand had failed, would we have avoided World War I?
If the Germans hadn’t sent Lenin in a sealed train car to the Finland Station, would the Bolshevik Revolution have happened?
If the United States hadn’t dropped the atomic bombs, would it have had to invade Japan?
If Thurgood Marshall had retired in 1980 rather than in 1991, would there have been a liberal majority for Bush vs. Gore and most of the 2000s?
If Al Gore had been president in 2003 instead of George W. Bush, would the United States have invaded Iraq?
If Hillary Clinton had spent more time in Wisconsin and Michigan, would we have been talking about her popular vote landslide in 2016?
These are all reasonable questions that strike at deeper themes. They force to you interrogate evidence and isolate the decisions that truly mattered versus those that were epiphenomenal or irrelevant. This is a valuable exercise, especially because what seems important in the moment may actually be of transient significance in the long-run and what seems marginal might have vast significance.
The trick, when not writing speculative fiction, is to put up some guardrails. Philip Tetlock and Aaron Belkin explored some potential ideas in their edited volume on the subject in 1996. Niall Ferguson articulated some of his own in the introduction to Virtual History (2000).
Three tests that make the most sense to me. To be useful, counterfactuals must be:
Plausible and grounded (departing from evidence on alternative paths based on the choices historical actors considered at the time or contingencies that were possible and reasonable, if not necessarily actively discussed);
Limited (hinging upon a single variable — an illness, a storm, a missed communication, an electoral upset, a single decision — rather than simply altering the situation beyond recognition); and
Appropriately caveated (appropriately humble about what we can reasonably project once we get into the realm of second- and third-order consequences. One has to recognize the possibility of intervening factors and future contingency, rather than simply spinning a pleasing narrative.
These rules separate the “parlor game” quality of the counterfactual-as-historical-fiction from its practical utility: to historians, to citizens, and to policymakers seeking a better understanding of how to make decisions. To return to Yglesias’ TR counterfactual, it passes the “limited” test, probably fails the “plausible and grounded” test, and certainly fails the “appropriately caveated” test — but again, interrogating it usefully surfaces these dynamics.
Recent Historical Work Using Counterfactuals
And in fact, there are a number of historians (popular and otherwise) who are producing valuable work that employs counterfactual history in one fashion or another.
To offer just a few examples:
Jim Steinberg, dean of Johns Hopkins SAIS and former Deputy Secretary of State under the Obama administration, wrote a fascinating article in the Texas National Security Review in 2019 exploring whether the decline in U.S.-China relations could have been avoided by looking at the policy choices between Tiananmen Square and the election of Donald Trump with reference to the decisions around Tienanmen, WTO accession, and the Scarborough Shoal crisis of 2012.
Philip Zelikow’s The Road Less Traveled: The Secret Battle to End of the Great War, 1916-1917 (2021) explicitly looks at how Wilson’s peace initiative, which could have worked, might have averted the Bolshevik takeover, the rise of Nazism, and the course of the twentieth century.
An edited volume entitled Historically Inevitable? Turning Points of the Russian Revolution (2016) draws on a range of superlative scholars of Russia (including Dominic Lieven, whose essay on potential foreign interventions from 1905 to 1920 is a must-read) to examine whether the Revolution might have been averted.
Frank Harvey’s Explaining The Iraq War: Counterfactual Theory, Logic and Evidence (2011) uses an explicit counterfactual method to argue that Gore would have gone to war had he been president due to the failure to contain Iraq in the 1990s and the broad bipartisan consensus favoring regime change in Iraq, explored at greater length by Melvyn Leffler, Joseph Stieb, and, most recently, Steve Coll.
And farther back, one can look to the aforementioned Niall Ferguson-edited Virtual History: Alternatives and Counterfactuals (2000) or the Robert Cowley-edited What If? series for provocative essays of varying quality.
It is a reflection of the state of academia that most of the historians that make these contributions are either working out of policy schools in the United States (like my home institution of SAIS), at certain universities in the United Kingdom, or as independent journalists and scholars.
It is not exactly correct to say, as Yglesias does, that historians won’t step up to ask these questions — but he is right to suggest that it isn’t the kind of thing that will get you tenure if you engage in it too extensively or vigorously, and that it is probably less common than it should be. We don’t need to enforce counterfactual history in the academy, but its value should extend beyond the limited environments where it thrives.
What sets all of the aforementioned contributions apart from less rigorous counterfactual history is that most of the writers above are judicious enough to follow the criteria above rather than embracing their inner alternate history novelists. You can profitably explore a counterfactual by isolating an event, exploring its context, and looking for clues in what people thought and did as well as purely material facts. The ratio of fact to speculation decidedly favors the former, even when the questions are driven by the latter: could things have been different? That question opens up a range of possibilities — and the quirks of what people were focused upon — are often richer and stranger than one might initially suspect, and reveal “unspoken assumptions” that might otherwise elude us.
The key is to embrace bounded speculation that doesn’t overtake the substance of the past. The historian or journalist who engages in responsible counterfactual history allows their imagination to wander within a well-tended walled garden that keeps the wilder outgrowths carefully pruned and in proportion to orderly beds. The walls should be low to allow a view from a commanding vista of the growth in the. surrounding hills — but in contrast to alternate history novelists, the responsible non-fiction writer’s subject is imaginative but necessarily constrained.
The Counterfactual and the Policymaker
Counterfactual reasoning is therefore not merely idle speculation or an antiquarian exercise. It can have direct relevance for policymakers when employed judiciously and can be viewed as the retrospective equivalent of scenario-based planning, which also focuses on isolating key variables, anchoring them to probabilities, and projecting potential outcomes.
There are interesting examples of efforts to use counterfactual history to examine a contemporary policy challenge in my experience dealt with the question of whether it would have been possible to prevent the Syrian Civil War from becoming such an appalling human tragedy.
In 2017, the United States Holocaust Memorial Museum released a series of studies that examined key moments in U.S. policy towards Syria, including two papers by Mona Yacoubian and Daniel Solomon that used counterfactual analysis. The study itself generated a great deal of controversy, in part because some of the conclusions suggested that little could have been done to coerce Assad given the asymmetric stakes for his regime and that there were tradeoffs in trying to remove him from power versus ending the conflict.4
But it asked relevant questions about an ongoing conflict — and highlighted core dilemmas by looking back at past decisions that had emerged from live and consequential policy deliberations. Considering that the war continued to rage when the study came out, the relevance of such an exercise should be obvious.
More broadly, though, the study of history, awakens the mind to all sorts of possibilities, connections, and constraints in the past and the present. That study should be sharped by counterfactuals, provided that they are used judiciously and rigorously.
Incidentally, Fatherland was the first book I checked out of Widener Library during freshman year, beginning a long pattern of undergraduate reading what I wanted rather than what as assigned. It was adapted by HBO into an adequate 1994 film starring Rutger Hauer and Miranda Richardson. (Those interested in nipping such an outcome in the bud should try another recent alternate history exercise — the text-based game Social Democracy, in which the player tries to keep the Weimar Republic afloat as the leader of the SPD.)
Evans’ book Lying About Hitler (2002) offers an excellent account of how he went about his task of demonstrating that Irving’s conduct was intentional, irresponsible, and mendacious. It's a forensic view of how historians’ use their methodology to assess difficult claims.
The counterfactual of reversing party racial realignment later in the twentieth century is intriguing, however. A George Romney candidacy in 1968 (given Romney’s support for open housing and active courtship of African Americans in Michigan) might be a more compelling case.