The US military’s remote warfare has recklessly killed hundreds more civilians than previously disclosed, including many children, as we’ve learned from incredible reporting by the New York Times.  What the Times didn’t say is that modern moral philosophy has been complicit in this moral calamity, has indeed laid the intellectual groundwork for it.  The fault lies in a trend in philosophy that reduces complex moral and political questions to reductive, God’s-eye judgements about who is “liable to be killed.”  The culprit is an approach I’ll call “Trolleyology” and its extension into the ethics of war.

I use “Trolleyology” to refer to philosophy that relies on stylized examples – many of them rather preciously involving trolleys – of choices whether to save or kill different numbers of people.  It has become a philosophical phenomenon, with at least two recent popular books exploring the questions, and routine news appearances.[i]  The core of the problem is explaining why, if it’s all right for someone to divert a runaway trolley from a track where it will kill five people, to a different track where it will kill only one, it is nonetheless wrong for a surgeon who can save five people dying of organ failure to do so by killing and harvesting the organs of a healthy person in for a checkup.  Philosophers then ring variations on these themes, for instance stopping the train by pushing someone else onto the tracks or having a bystander rather than a driver pull the switch, with the aim of extracting general principles that regulate when it’s okay to divert threats even at the cost of others’ lives.

Members of the general public who have heard of Trolleyology probably think its puzzles are amusing and harmless.  Within philosophy, I suspect most think of it as empty because it reflects actual ethical choices so poorly: these rarely involve immediate life or death, virtually never include the certainties of the outlandish thought experiments, and always involve important issues of context, relationships, and community.[ii]  But emptiness is the least of it.  The inane examples of Trolley Problem ethics conceal a radicalism about individual rights and culpability that, though irrelevant in daily life, has nourished the practice of remote war and targeted killing.

Trolleyology originated almost half a century ago in 1967 article by the Oxford philosopher Phillipa Foot, “The Problem of Abortion and the Doctrine of Double Effect,” where Foot laid out the case for some forms of abortion by starting with the strongest possible anti-abortion premise, that a fetus has the same rights as an adult.  She then considered under what conditions it might still be permissible to cause the fetus’s death in order to save the mother’s life.[iii]  Foot therefore thought up examples that put adult lives in conflict, such as when they’re on alternative trolley tracks or blocking each other’s exits from caves.  She argued that while we can choose to kill fewer people rather than more if we must kill either way, we cannot choose to kill some people rather than none to prevent more deaths.  She explained the special wrongness of killing by distinguishing between stronger “negative” duties not to interfere with someone’s life and “positive” duties to help them; rights to be saved, she said, were less stringent than rights not to be harmed.  Applied to abortion, her conclusion is nearly as restrictive as Catholic doctrine: one cannot destroy a viable fetus even to prevent the mother’s death, although one can (contrary to Catholic doctrine) intentionally destroy a fetus to save the mother if otherwise both would die.[iv]  Her mostly anti-abortion conclusion follows from a logic that abstracts the question of abortion from all aspects of the situation of gestation and pregnancy, which one might think would render the whole discussion irrelevant.  Is it OK to eat oysters?  What if the oyster you’re eating were actually the last remaining member of an almost-extinct species of bird?  What if you’d starve to death otherwise?  What if you wouldn’t actually starve to death, but eating the almost-extinct species of bird was the only thing that could save you from progressive blindness?  What if this particular bird was about to die before reproducing anyhow, so the species was inevitably going to be extinct?  So then, it’s OK to eat oysters but only in certain very rare circumstances?

The next stop on Trolleyology’s route was the MIT philosopher Judith Jarvis Thomson’s 1971 “A Defense of Abortion.”[v]  Thomson, who would go on to author three canonical treatments of the Trolley problem (which she also named), resorted to an even more outlandish example: you wake up to discover yourself connected to a concert violinist, who needs for some reason to live off your kidney for the next nine months.  Of course, she says, you can detach the violinist.  Thomson’s modest conclusion, permitting abortion in cases of rape, rebuts only extreme anti-abortion positions.  Like Foot, she ignores the broader context of pregnancy and abortion – including how they bear upon questions of social and economic equality – reducing the issue to a question of individual rights.  Detaching the parasitic violinist is self-defense, grounded in the host’s right to an autonomous existence, even against someone not at all responsible for the threat they pose.  She makes the right to abortion, at least for cases of rape, look like a right not to provide lifesaving aid to an adult.  It is, in fact, a position of extreme libertarianism that begins with how to protect ourselves against the claims of others.

The libertarianism of trolley morality becomes clear in the work of the Harvard philosopher Robert Nozick, where its focus shifts from feminism to capitalism.  His 1974 Anarchy, State and Utopia is a touchstone for economic libertarians for its argument against redistributive taxation.[vi]  The work has become at least as influential in discussions of self-defense.  Nozick invented two examples to argue that individuals may defend themselves with lethal force against “innocent aggressors” (non-culpable threats) and at the cost of the lives of “innocent shields” (passive third parties who will be harmed by the self-defense).  His examples are in the outlandish, cartoonishly violent Trolleyology tradition and involve people thrown down wells, whom you (at the bottom of the well) may vaporize with your “ray gun,” and others strapped to the fronts of attacking tanks.[vii]  The examples force an urgent choice: between them and me, may I save myself?

Scenarios involving tanks and wells support Nozick’s central example of “Wilt Chamberlain” who gets rich through fans voluntarily paying a fee to see him play.  This example is the mainstay of his argument against redistributive taxation, on the grounds that a tax system would require a wrongful “continual interference” with the preferences of the would-be spectators by preventing them from paying to watch Wilt play (presumably because if he’s taxed he’ll decline to play?) .  The social interests that would be supported by redistribution are like people falling down wells: you have no duty to suffer them.  Once again, Nozick’s trolley logic ignores all contextual considerations: the economy that provides spectators with their incomes, the political and social costs of wealth inequality, the automatic, background character of modern tax systems. The Trolleyological trick is to load into the example only normative features that favor the desired conclusion.

Since morality is not just socially embedded but the fabric of social life itself, you might wonder why philosophers would find deracinated examples useful as a way of discovering fundamental moral truths.  Here another aspect of Trolleyology is relevant: its scientism.  By “scientism” I mean reasoning that adopts the forms of science as a way of achieving intellectual respectability.  Scientism involves a kind of professionalization, including spurious use of formal notation, a claim to view-from-nowhere objectivity, and the proliferation of technical jargon – all familiar aspects of modern philosophy.  It is also scientistic to suggest that we can discover moral truths through an experimental, rather than discursive, method.

Today’s leading Trolleyologist, the Rutgers philosopher Frances Kamm, argues that our subjective intuitions in bizarre examples are guides to an objective morality, despite the fact that individuals disagree with each other and even with themselves over time. (Thomson, for instance, has radically changed her views at least twice.)[viii]  At the apex of moral scientism is the Georgetown legal scholar John Mikhail, who has set out to show that patterns of moral judgment are universal. [ix]  He attributes these patterns to a moral analogue of the innate grammar faculty that Noam Chomsky posits to explain how any healthy baby can come to speak any human language.  So too, Mikhail says, can anyone arrive at the same answer regarding when someone may divert a rampaging trolley.

Mikhail aims to demonstrate this universal moral intuition by means of global opinion surveys.  But the project quickly founders.  Except for the two widely-agreed poles of Trolleyology –a bystander can divert the train to kill one saving five, and no one should stop it by throwing a bystander under its wheels – most responses indicate ambivalence, not universal agreement.[x]  People generally share basic values of preferring less harm to more and avoiding becoming the direct source of harm to others; but balancing these two values is often hard.

Trolleyology’s mistake is reduction: the illusion that the surface level complexity of some phenomenon masks a deeper simplicity, an atomic vocabulary of rights, intentions, and causes.  Reductions may be possible in some areas of science, where complex phenomena might be explained by the lawlike behavior of some set of simpler properties.  But ethical and political situations are not reducible because they are the complexity.  Complexity is their very essence.

There are not yet automatic systems in civilian life implementing the results of trolley problem surveys.  However, one context has lent itself to trolley-type examples: military decisions about how to balance deaths of soldiers and civilians in pursuit of military advantage.  Nozick used Trolleyology to argue against taxation and in favor of letting people die of penury.  Recent uses of Trolleyology go a step further, authorizing killing people outright.  Both are the results of an abstract, reductive theory of individual rights.

These recent uses of Trolleyology to justify killing the innocent come as part of “the new Just War Theory,” a wave of work among moral philosophers in response to America’s Iraq and Afghanistan wars, attempting to determine when states can go to war and what means they can use.[xi]  Traditional Just War view is heavily state-centered: only states have a right to use military force, and only when their territories or peoples are threatened.  But individual soldiers have a right to use force so long as they are ordered to fight by their states, whether their state is an aggressor or victim.

Traditional Just War theory also distinguishes between combatants whom soldiers can target deliberately, and non-combatants whom soldiers cannot attack directly.  Soldiers may knowingly kill non-combatants if they don’t directly intend their deaths, but their deaths must be a necessary and proportional effect of an attack on a legitimate target.  The distinction between combatants and non-combatants is a matter of status: soldiers can target enemy combatants even when they are not carrying weapons, and even if they are coerced draftees, while soldiers must protect civilians, even if they are working in war factories, or voting for and paying taxes to militarist governments.[xii]

These fundamental principles may appear corrupt since they protect the civilian politicians who order wars while making vulnerable the draftees who fight them.  They also permit a great deal of killing of combatants and non-combatants alike.  But the alternative to these principles is usually thought to be no principles – the kind of total war exemplified by World War II, in which militaries killed more far more civilians than soldiers, many from aerial bombing.[xiii]

In the last two decades, philosophers influenced by Thomson and Nozick have used their approach to rewrite the traditional Just War principles.  These “Revisionists” propose to look at the problem of war through the atomistic lens of individual morality instead of the collective, state-based lens. [xiv]  Since ordinary morality, mainly concerned with promise-keeping, lying, and duties to help strangers, has had little in to say about killing except for an injunction not to do it, the main resources available to Revisionists have been moralized versions of the criminal law governing individual self-defense – and Trolleyology.

In pursuing their Trolleyological analysis of war, Revisionists have bypassed any idea of a state.  Their working hypothesis is that they can analyze large-scale conflict into a web of individual acts of attack and defense, some justified, some unjustified. War is a larger version of an armed robbery met by police, where the question is who has a right to shoot whom.  In the language of the Revisionists, the robbers are “liable” to be killed: people who kill them do no wrong.

One can be skeptical of the traditional Just War theory because of its permissiveness towards mass violence.  But at least the tradition takes war seriously as a collective, political endeavor, treating it as the object and context of analysis.  The revisionist approach, reducing war to a complex of individual altercations, founds itself in moral “intuitions” that are rooted in criminal laws governing individual self-defense, without acknowledging that these presuppose historically specific versions of state authority and background institutions for resolving conflicts.

Some critics of the revisionist Just War theory have argued that individual self-defense claims do not apply well to war, since ordinary civilians are rarely subject to a lethal threat that they could only ward off by violence.[xv]  This seems to me a persuasive criticism.  But I think the original sin of Trolleyology is its method of reducing complex social and political questions to simplistic, outlandish, contrived examples.  These examples assign the causal role of each individual and present all the relevant information in advance.  Then they ask you to decide how you would weigh your life against someone who will harm you if you do nothing.  Such reasoning-by-example leads the philosopher to a clear conclusion: if only one can survive, it’ll be me.

The next step extends to considering the liability of people indirectly supporting threats, for instance an ambulance driver who saves a potential murderer who can then kill; a farmer whose crops feed soldiers in a war of aggression.  Revisionists use their contrived examples to endorse the killing of Red Cross workers who, aiding just and unjust combatants alike, unintentionally make further attacks likelier; or taxpayers, whose taxes support the army.[xvi]  Once we can identify someone’s causal support for a threat, it doesn’t matter whether it takes the form of a trigger finger, a vote for a militarist candidate, or money for weapons.  The analogy to a direct “you or me” scenario transforms the reasonable judgment that civilians supporting unjust conflict bear some moral responsibility, to the much harsher judgment that they therefore deserve to die. The result is a moral green light to attack civilians.

Revisionist Just War theory, founded in Trolleyology, ignores the social and political context in which wars are conceived and fought, including the deeply-rooted distinction between soldiers and civilians, the effects of propaganda and informal coercion, and the ways in which seeing whole populations as deadly enemies jeopardizes any future peace.  We should also notice that the popularity of a reductive, hyper-individualized ethics of war coincides with the modern military practice of powerful states.  Trolleyology has become a required subject of study at West Point, where it teaches soldiers to think that combat can be mapped onto cute moral schemas of killing.[xvii]  The current situation of endless conflicts involving targeted kill squads and drone strikes finds its spurious moral vindication in the reduction of war to individually-targeted retributive justice.

Trolleyology, with its pristinely abstract world of reductive scenarios, has given rise to the illusion of the morally costless war.  Look at the recent drone killing by the US of ten Afghans, seven of them children, erroneously thought to be linked to a terrorist threat but who in fact were relief workers and their families.  Such killings happen because generals promise that war can be waged “over the horizon” and that they can make just decisions to kill with gods-eye certainties about who is “liable to be killed” and whether the numbers of the rightly killed offset the unintended casualties.[xviii]  The dream of a computable ethics, reducing problems of life, death, and politics to a universal calculus, has done terrible harm.  It is a mad dream, since ethics does not detach from life.  Ethical choice will always be as messy as the contexts in which it arises.  Defenders of drone programs insist that they are better than the ground forces whose record will be far worse.[xix]  But in fact, drone strikes rarely take the place of military occupations.  Instead, they supplant non-violent means of maintaining peace such as diplomacy and aid for countries’ internal political capacities.

Ethics is not a science.  It is a difficult conversation, filled in by many voices in chorus and dissent.  In resolving what to do about the daunting problems confronting us, from armed conflict to climate change to undoing the wrongs of an enslaving and colonial past, we cannot be guided by simplistic scenarios that atomize moral situations into individual bits.  Nor should we overlook the libertarian politics underlying these, with its tendency to justify self-serving violence.  For all its air of quaint charm, Trolleyology has served sinister purposes.



[i] See David Edmonds, Would You Kill the Fat Man? (Princeton: Princeton University Press, 2010); Thomas Cathcart, The Trolley Problem; or Would You Throw the Fat Guy off the Bridge? (New York: Workman, 2013).

[ii] As explored by James Wilson, “The Trolley Problem Problem,” Aeon 28 May 2020.

[iii] Oxford Review 5 (1967).

[iv] Catholics held that this was only permissible in the special case of a cancerous uterus, whose removal would indirectly cause fetal death.  Later Foot says that until the question of the moral status of the fetus is settled, nothing definitive can be said about abortion. Moral Dilemmas and Other Topics in Moral Philosophy (New York: Oxford, 2002), 87.

[v] Philosophy and Public Affairs 1: 47-66 (1971).

[vi] Robert Nozick, Anarchy, State and Utopia (New York: Basic Books, 1974).

[vii] Nozick, 35.

[viii] She first stated her view “The Trolley Problem,” Yale Law Journal 94: 1395-1415 (1985), then changed her mind in The Realm of Rights (Cambridge: Harvard, 1990), and again in Turning the Trolley.” Philosophy and Public Affairs 36, no. 4 (Fall 2008): 359–74.

[ix] John Mikhail, Elements of Moral Cognition: Rawls’ Linguistic Analogy and the Cognitive Science of Moral and Legal Judgment (New York: Cambridge University Press, 2011).

[x] Mikhail, “Universal Moral Grammar: Theory, Evidence, and the Future,” Trends in Cognitive Science 11, No. 4: 143-151 (2007).

[xi] The traditional Just War view is explained in Michael Walzer’s Just and Unjust Wars (New York: Basic Books, 1977).

[xii] According to the traditional view, codified as Rule No. 1 of customary international humanitarian law, soldiers may target non-uniformed people who directly engage in combat; the hard question relates to civilians who play direct but occasional roles in conflict, for instance serving as spotters.

[xiii] Estimates vary, but a credible one is that nearly 30 million civilians died from direct military causes, compared to 23 million combatant deaths.  At least another 20 million civilians perished from famine and disease.  “World War II Casualties by Country,” World Population Review.

[xiv] See, e.g., Jeff McMahan, Killing in War (New York: Oxford, 2011); Cécile Fabre, Cosmopolitan War (New York, Oxford, 2012); Helen Frowe, Defensive Killing (New York: Oxford, 2014).  Seth Lazar has an overview of the debates, in “Just War Theory: Revisionists and Traditionalists,” Annual Review of Political Science 2017: 20: 37–54.

[xv] David Rodin, War and Self-Defense (New York: Oxford, 2003).

[xvi] Frowe, Defensive Killing, 210.

[xvii] Edmonds, “Lessons in Morality at West Point,” BCC Radio (18 September 2010).

[xviii] Trolleyologists note the uncertainty about who will be saved by the decision to kill, but then ignore it and treat the problems as a good proxy for wartime decisions.  See, e.g., Markus Christen et al., “Trolley dilemma in the sky: Context matters when civilians and cadets make remotely piloted aircraft decisions,” PLoS ONE 16(3): e0247273 (2021).

[xix] See, e.g., Daniel Statman, “On Drones and Robots: On the Changing Practice of War,” The Oxford Handbook of War, ed, Seth Lazar and Helen Frowe (New York: Oxford, 2018).