At a time of renewed debate in the United States over the merits of using coercive interrogation techniques by intelligence services and law enforcement, John Schiemann’s book is an important reminder of the inherent limits of torture. What is all the more remarkable is that Schiemann does not rely on the moral or legal critiques by various philosophers or legal scholars, but rather approaches the claims put forward by torture proponents on their own terms. The book’s opening epigraph from human rights scholar Michael Ignatieff rightly captures the tenor throughout this systematic treatment of such a charged topic: “If we are against torture, we are committed actually to arguing with our fellow citizens, not treating those who defend torture as moral monsters.”

The main foil throughout the book remains the torture regime, particularly the CIA program of the George W. Bush administration during the War on Terror. Yet rather than portraying the Bush regime as some aberration of abusive excess, Schiemann makes a compelling case that the U.S. program designed during this period closely approximates the “idealized” model envisioned by more pragmatic proponents of how a torture regime—properly formulated, regulated, and implemented—should be expected to function. Assessing the Bush program thus offers a more general test of any torture regime, however ideally conceived, with implications for supporters and opponents alike.

The opening chapters lay out in clear, oftentimes brutally stark, terms the defining features of torture, alongside the conditions under which proponents would judge an interrogational program employing torture as successful on pragmatic grounds. Weaving together existing scholarly work with primary documents, Schiemann distills benchmarks upon which torture programs should be assessed based on four key elements: 1) high reliability of the information collected; 2) low frequency of torturing only the most resistant of detainees; 3) limiting the severity of abuses; and 4) the success of such calibrated torture in getting resistant suspects to give up most (if not all) information while sparing cooperative or innocent detainees. The remainder of the book goes on to demonstrate in detail how none of these benchmarks can reasonably be met, even under the most ideal of circumstances favorable to proponents’ positions. That Schiemann finds the Bush program wanting across multiple dimensions offers an even more fundamental challenge to scholars, policymakers, or members of the public who believe that torture represents an effective policy under almost any circumstances.

Does Torture Work? builds its case with a game-theoretic model of interrogation. Schiemann starts with relatively simple building blocks of a single-shot sequential game involving two players—a detainee and an interrogator. Each has two choices, respectively, whereby the detainee either chooses or refuses to give information, and the interrogator then decides whether or not to use torture to try to get the detainee to talk. Such a straightforward model with fully informed players allowing slight differences in preference orderings yields puzzling (or, for Schiemann, “quixotic”) outcomes, such as torture never happening or, instead, information never being obtained.

Given that neither equilibria mesh well with what is known from real-world episodes, Schiemann goes on to incorporate insights from historical and contemporary cases to develop a more realistic interrogation model (RIT). Keeping the same overall structure, an incomplete information game is developed to introduce uncertainty over the particular “type” of actor both sides believe they are facing—whether the interrogator is pragmatic or sadistic; and whether the detainee is cooperative, resistant, or innocent. The model is also flexible enough to allow for different types of questioning—leading versus objective—to capture important differences in interrogation scenarios. While necessarily simplifying the types of actors and actions, the RIT does what any good formal model should do in highlighting some of the most salient dynamics involved in complex phenomena. Throughout these formative sections of the book, Schiemann describes and justifies each modeling choice in clear terms, providing further discussion and examples where necessary. For this alone, the book serves as an excellent template for how to present and discuss a formal model to a general audience.

The modifications in the RIT turn out to be crucial, as solving the model generates a dizzying nine separate equilibrium outcomes. Schiemann deftly organizes these varying solutions into a common framework dependent upon three core beliefs—whether the interrogator is pragmatic, whether the detainee revealed all of his or her information, and whether a detainee is innocent if no information is provided. The rest of the chapters investigate each equilibrium in detail. Brief historical narratives illustrate the results, finishing with a lengthier postscript discussing insights from the 2014 Senate report on the CIA program. But ultimately, the game-theoretic model itself represents the main evidence marshaled by Schiemann.

Indeed, one of the broader and more explicit aims of this work is to argue that mathematical modeling, largely on its own, can inform important questions of political science and philosophy, especially where data issues are challenging, as with interrogation. The author exploits the model to bring together a number of propositions and implications that question every benchmark of any idealized vision of a torture regime. He shows at each turn that torture is unlikely to yield reliable information, will be more frequent and severe than reasonably acceptable, and engulfs a disproportionate number of innocent and cooperative victims. Taken as a whole, a convincing case is made for the limits of torture, and the book’s unique approach complements other more theoretical and historical works, such as those by Darius Rejali (Torture and Democracy, 2007) or by Karen J. Greenberg and Joshua L. Dratel (The Torture Papers, 2005).

The heavy lifting done by game theory in Does Torture Work? raises a number of questions and concerns, both internal and external, to this approach. Given that a formal model is only as insightful as its underlying modeling choices, there are a number that deserve further inquiry, especially as they do not always match up well with the illustrative cases. In particular, the full costs to the interrogator of engaging in torture are never fully elaborated or incorporated. The model acknowledges that interrogators may incur some psychological or morale consequences (among others) from engaging in torture, but assumes that these are minimal and are at best less than those of not torturing a reluctant detainee. One of the book’s own cases strongly suggests otherwise. The episode involved member after member of the Frankfurt police force in Germany refusing to torture a kidnapping suspect until one volunteer had to be helicoptered in to do so. The reasons given, largely focusing on officers’ commitment to the law, point to a dimension not easily captured in the book—that torture occupies a particularly heinous place in both domestic and international law, and can generate distinctive normative aversions beyond the game’s more straightforward cost–benefit analysis.

The costs are likely even greater when a longer view is taken. This is evident even in the book’s focal case of the Bush program. After reports of torture surfaced, the global image of the United States was certainly hurt, but this was compounded by a chilling of substantive counterterrorism cooperation from other countries, including intelligence sharing, extradition, and basing agreements (Emilie M. Hafner-Burton and Jacob N. Shapiro, “Tortured Relations: Human Rights Abuses and Counterterrorism Cooperation,” PS: Political Science & Politics, 59(3), 2010). This relates to another restrictive assumption of interrogation as a single-shot game. Both the Bush regime and other cases instead show an inherently repetitive process where both history and the future matter beyond the fate of the immediate detainee. The eponymous “Nature” is unlikely to be the only factor deciding the types of actors in an interrogational game of incomplete information. Rather, prior decisions involving torture may endogenously affect each side’s evaluation of the other’s motives, and ultimately even the frequency and possibility of future interrogations. One legacy of the Bush years was that the Obama administration largely switched to targeted killings, avoiding the pitfalls of interrogation and detention, but lost valuable opportunities of intelligence gathering as a result. This all potentially leads to an even stronger conclusion than Schiemann’s—not only that torture is no more effective than more humane practices but furthermore, that it may also be fundamentally counterproductive.

This points to a second, more general question about the author’s great reliance on game theory for drawing inferences. Reading this book brought back reminders of other works using game-theoretic models of varying sophistication to study similar topics, such as Robert Axelrod’s use of the iterated prisoner’s dilemma to explain World War I’s live-and-let-live system (The Evolution of Cooperation, 1984), or Niall Ferguson’s similar reliance to account for surrender rates during the two world wars (“Prisoner Taking and Prisoner Killing in the Age of Total War,” War in History, 11(2) 2004). Yet both works have been criticized for imposing restrictive models of strategic interaction on events that could better be explained by a closer and contextualized treatment of the underlying actors and dynamics at play. (See Brian Dollery and Craig R. Parsons, “Prisoner Taking and Prisoner Killing: A Comment,” War in History, 14(4) 2007; Andrew Gelman, “Methodology as Ideology: Mathematical Modeling of Trench Warfare,” QA–Rivista dell’Associazione Rossi-Doria 2, 2007, 167–176). To borrow from another of the book’s opening quotes, here from John Arbuthnot, mathematical reasoning can serve as a candle to illuminate the darkness, but it may tend to shine on those elements more easily modeled to the detriment of obtaining a fuller and more convincing picture.

These caveats notwithstanding, Does Torture Work? is a thought-provoking work that generates novel insights and avenues on a controversial topic. It should be required reading for anyone interested in the forces at play in the continuing, and in many ways growing, debate over torture.