From consequentialism to deontology

ILPI Publications > Statements

Torture, hostage-taking, and nuclear weapons

By Nobuo Hayashi
15 September 2015

Presented at “Humanitarian Impact: Why Ethics Is Important to the Politics of Nuclear Weapons,” a side-event held during the 2015 Non-Proliferation Treaty Review Conference.

2015 NPT Review Conference Published: September 2015


My aim here is to suggest a change in the angle from which we consider the ethics of nuclear weapons. We should reject the security-focused, consequentialist premises on which both moral arguments for and against nuclear weapons have traditionally been built.

We have been asking ourselves whether nuclear weapons can be good for something. To this, the weapons’ proponents would say: “Yes, they are good for national security, self-defence, and international stability.” To the same question, many of us who oppose nuclear weapons would reply: “No, they undermine these things.”

I suggest that we have all been asking the wrong question. We should reframe the question to: “Should it matter whether nuclear weapons might be good for something?” And we should respond to this rephrased question with: “No, it need not matter.” Or, better still: “No, it ought not matter.”

Using and threatening nuclear weapons are best seen for what they really do. The moral status of using or threatening these weapons is more appropriately assessed on the unparalleled suffering they are bound to inflict, rather than what purpose their threat or use supposedly serves.

Just as we consider torturing and taking hostages unjustifiable under any circumstances, using and threatening nuclear weapons should be deemed fundamentally unethical.

The real challenge ahead is whether we can overcome our own consequentialist mentality, and generate a broad political consensus that nuclear weapons’ singular inhumanity makes it inherently unethical to use or threaten them, and that it is so whatever their real or imagined utility. This year’s NPT Review Conference is an important occasion to nurture such a consensus.

Asking the wrong question

Why is “Can nuclear weapons be good for something” the wrong question to ask? Consequentialism does not merely look at what an act causes. No one seriously doubts that nuclear weapons cause horrendous consequences for their victims. These “consequences,” however, are not the relevant point of reference for this mode of ethical reasoning. By “consequences,” consequentialism looks at the purpose of an act, and treats this purpose as the ultimate ground for justifying the act.

To those who seek to legitimise nuclear weapons on consequentialist grounds, what matters is not that they cause massive harm, which they of course do. What matters to them is rather that the harm they cause is “worth it.” This harm is “worth it,” on this view, because nuclear weapons help with their possessors’ self-defence, national security, and so on.

The fact that nuclear weapons inflict unspeakable inhumanity – the subject of the three international conferences on humanitarian impacts recently held in Norway, Mexico, and Austria – is the very reason for which they are deemed so valuable. It is precisely because the harm nuclear weapons cause is so unacceptably high for their target, that these weapons can be useful as deterrents in the first place.

This is what makes nuclear weapons so ironic. Their singular inhumanity, and their real or imagined utility as deterrents, are but two sides of the same coin.


There are three reasons for which “Can nuclear weapons be good for something?” is the wrong question to ask.

First, this question is futile. It is futile, because neither consequentialist arguments for nuclear weapons, nor those against them, can ultimately be verified or falsified – involving, as they both do, counterfactuals, the what-ifs of alternative histories, and the maybes of rival futures.

In a nutshell, consequentialism asks us to compare two things. One is the actual level of international peace and security, with the actual level of global nuclear armament in it. The other is an imagined world with fewer, or no, nuclear weapons. Then we are to wonder how history would have unfolded, where we would be today, and what the future might hold in store for us.

For proponents of nuclear weapons, today’s nuclear world is the world that, though admittedly imperfect, has more or less shown itself to work. Empirically speaking, managing risks, ensuring that nuclear weapons remain in the hands of reliable players, and strengthening the status quo generally, may very well give us the best overall chance of continued security. Kenneth Waltz, a prominent neo-realist international relations scholar, even suggested: “With more nuclear states the world will have a promising future.”[1]

Opponents counter that nuclear weapons are not only evil but also detrimental to international peace and security. If anything, it is “despite” these weapons, not “thanks” to them, that we have kept our world more or less secure. Notwithstanding all our best and most prudent intentions, there is no guarantee that we will continue to hold the world together in the future.

The problem is that it is impossible to confirm claims of but-for causation that are implicit in consequentialist justifications. For instance, there is no way one can conclusively establish that Ukraine would not have been in this situation of difficulty had it kept its nuclear weapons. Nor, for that matter, is it possible to verify their counterfactual objections. Thus, claims such as that according to which Ukraine would not have had to experience emergencies like this one if we had not had nuclear weapons in the first place, can never be proven.

We simply cannot know whether greater or fewer numbers of nuclear warheads would have brought us more stability. Future predictions, however technically sophisticated, never really go beyond the realm of educated guesses. How can one know that a future world with or without nuclear weapons is also a more secure one?

Bias in Favour of any Workable Status Quo

Second, “Can nuclear weapons be good for something?” is the wrong question to ask, at least for their opponents. Our inability to prove the desirability of a world free of nuclear weapons means that any workable status quo enjoys a moral advantage over that world.

By rejecting nuclear weapons on consequentialist grounds, we would be essentially suggesting that we should choose to deal with the devil we do not know at all, rather than with the devil we know. This is an exceedingly tall order, and hardly a compelling moral choice. That is so, because we are up against a model of the world that has sustained itself tolerably well: For all its bumps and scary moments, it has so far managed to avoid another nuclear exchange.

No matter how deftly one highlights the flaws, dangers and risks of nuclear weapons – though this is something that the Humanitarian Initiative has done rather well – , doing so does not, in and of itself, validate the supposed superiority of their absence. A world free of nuclear weapons might suffer from some other, but just as deadly, threats to peace and security.

Evidence, Not Principle

Third, for those of us who oppose nuclear weapons, “Can these weapons be good for something?” is the wrong question for another reason as well. Opposing these weapons on consequentialist grounds is merely a matter of evidence, not one of principle.

We can certainly argue – perhaps even prove – that nuclear weapons are no good for international peace and stability for this or that reason. Even if we can do all this, we actually are not challenging the underlying consequentialist premise, namely, that it is possible that nuclear weapons can be good for something. We are simply disputing the evidence adduced by these weapons’ proponents.

Arguing consequentialism against consequentialism exposes you to being persuaded that, if nuclear weapons do contribute to world stability and peace, then these weapons are morally acceptable.

Asking the right question

What, then, is the right question to ask? I would suggest we ask ourselves: “Should it matter whether nuclear weapons might be good for something?” This question, I would submit, is a moral question that is at least as valid as, if not more appropriate than, the question we have been considering so far.

And, according to one long-standing mode of ethical reasoning known as deontology, the answer is: “No, it need not matter.” In fact, deontology goes one step further: “No, it ought not matter.”

Debating whether nuclear weapons keep us safe misses the point of their intrinsic moral status. We should see these weapons for what they really do to their victims, not what purposes they allegedly serve, or how important such purposes supposedly are. The morally relevant suffering here is that which nuclear weapons inflict, rather than the suffering that is necessary or unnecessary for this or that end, is it not?

From a strictly deontological point of view, the singular inhumanity to which these weapons are bound to subject their victims, thereby reducing them to the status of mere sacrificial instruments for the benefit of the rest of humanity, makes it inherently immoral to use or threaten them.

In order to illustrate what it really means to say this, I wish to propose two moral analogies. The first is an analogy between committing torture and using nuclear weapons. The second is an analogy between taking innocent persons hostage and threatening nuclear weapons.

Committing Torture and Using Nuclear Weapons

We categorically reject torture. Moreover, we do so, even where torturing one suspected ticking bomber might save thousands of innocent lives. Torture is a moral wrong in itself, and under no circumstances is it ever justified.

Torture’s inherent immorality remains the same – not only because it often does not work, but also even if it happens to produce the desired result in some situations. Our rejection of torture is independent of its utility or disutility. Michelle Farrell, author of a recent book on the prohibition of torture in exceptional circumstances, notes:

The claim that torture might be justified in ticking bomb situations is, fundamentally, a claim that certain individuals do not have the right to have rights and, thus, that the human can be reduced to a status other than human.[2]

Using nuclear weapons can be considered analogous in moral status to committing torture. Decimating population centres and leaving generations of survivors with horrific aftereffects of radiation, all done in the name of some allegedly worthier goals, such as restoring strategic parity between nuclear-armed adversaries and expediting war’s conclusion, violate the same fundamental moral principle upon which we condemn torture.

Taking Hostages and Threatening Nuclear Weapons

As regards threats of nuclear weapons, however, we need to draw an analogy with some other act. The intrinsic immorality of committing torture does not quite generate the intrinsic immorality of threatening torture. Otherwise, we would have to enter a convoluted realm of ethical reasoning of conditional intention (i.e., to torture) in relation to another action (i.e., the victim refusing to relent and speak). Moreover, we would have to derive the moral status of threatening to do something indirectly from the moral status of actually doing it.

Taking hostages comes in very handy here. By definition, threatening victims with harm is an integral part of hostage-taking. Deliberate risk creation is an act that can be assessed on its own. Its moral status is tied neither to that of conditionally intending to harm hostages, nor to that of actually harming them.

Now, virtually all of us reject the idea that hostage-taking can ever be justified. Our unreserved condemnation of hostage-taking is independent of the act’s real or imagined utility, or whether or not the hostage-taker is entitled to demand the outcome. Hostage-taking is indefensible because, in the words of Steven P. Lee, a noted moral philosopher on nuclear deterrence, its victims

are treated as mere means in the sense that their holders use them by imposing a risk of harm on them in pursuit of the holders’ own ends. But what makes it invariably the case that hostages are treated as mere means is the fact that they are innocent and have not given consent.[3]

Similarly, nuclear deterrence amounts to taking entire populations, not just of one’s enemy but also of third states, hostage. Nuclear-armed states impermissibly use the inevitability of their unspeakable suffering as a “collateral” against future attacks. “[N]uclear deterrence cannot be justified by the fact that the behaviour the threatening nation seeks from its opponent – nonaggression – is something to which the nation is morally entitled,”[4] writes Lee.

Two challenges

Deontology, however, is not without difficulties. First, it awkwardly commits its adherents to sacrificing themselves on the altar of an absolute rule. Can we really insist that we accept our own demise, rather than act immorally by using or threatening nuclear weapons? Second, while deontological reasoning can stand on its own, its actual acceptance by society is more contingent on the historical evolution of our moral compass than on its logical vigour.

Rule Fetishism

There is no such thing as expedient deontology. In his exchange with another philosopher, Immanuel Kant famously argued that lying is inherently immoral. He stood by the notion that “it would be a crime to tell a lie to a murderer who asked whether our friend who is being pursued by the murderer had taken refuge in our house.”[5]

Kant might at least be given credit for being consistent. It is doubtful, however, whether such radical steadfastness would accord with our ordinary sense of a lie’s immorality in similar circumstances. Some even suggest that only “rule fetishists”[6] would go that far. Obsessing with the deontological unacceptability of lying comes at the expense of broader contexts and other weighty values.

Asserting nuclear weapons’ absolute immorality would amount to demanding that, should the choice be between self-destruction and recourse to these weapons, a nuclear-armed state choose self-destruction. This kind of rigid – one might even say suicidal – fidelity to moral consistency may appeal to some abstract forms of idealism. But it may not be suitable as a basis of responsible and pragmatic national policy. While deontology is good at highlighting salient ethical problems of an act, it is sometimes a mistake to disregard other important considerations.

Evolution of Our Moral Compass

This does not mean, however, that the deontological unacceptability of an act never gives rise to its actual prohibition. Our condemnation of torture did not emerge overnight. It has evolved gradually and intricately. In medieval times, torture was tolerated as an unpleasant yet necessary tool of justice and state power. It is the subsequent refinement of statecraft that rendered torture dispensable. Meanwhile, Enlightenment progressivism embraced human dignity as its centrepiece. Only then did a truly robust moral case against torture gather momentum and lead to its unqualified prohibition that we now take as self-evident.

Today, we are at a stage where our deontological conviction against torture is largely immune to accusations of rule fetishism. The same is true, mutatis mutandis, of our condemnation of hostage-taking.

A similar change regarding nuclear weapons is arguably afoot. The end of the Cold War has diminished the weight of its once all-encompassing logic and, with it, the perceived strategic value of nuclear weapons. The three aforementioned conferences held in Norway, Mexico and Austria have gathered mounting scientific evidence on the unimaginable suffering to which these weapons are bound to subject their victims.

These changes in our moral landscape should allow us to see nuclear weapons for what they really do, rather than what purpose they serve. We should reject nuclear weapons, not because they fail to serve the purposes that their proponents say they do, but because their use and threat are inherently immoral.


We should refuse to entertain the question: “Can nuclear weapons be good for something?” Rather, we should insist: “It does not matter if they might be. They are fundamentally unethical, whatever their real or imagined utility.”

We must insist on the elimination and prohibition of nuclear weapons at this NPT Review Conference, and whenever possible thereafter. We do so because, much like torture and hostage-taking, neither their use nor their threat can ever be justified under any circumstances. This is a moral imperative, regardless of the security situations in which we fin

This talk is partly based on a paper on the ethics of nuclear weapons, written by Mr. Hayashi for the ILPI-UNIDIR NPT Review Conference Series in 2015. Click here to access the paper.

[1] Kenneth Waltz, “The Spread of Nuclear Weapons: More May Be Better,” Adelphi Papers, vol. 21, no. 171, 1981.

[2] Michelle Farrell, The Prohibition of Torture in Exceptional Circumstances, Cambridge University Press, 2013, p. 246.

[3] Steven P. Lee, Morality, Prudence, and Nuclear Weapons, Cambridge University Press, 1993, pp. 45-46.

[4] Ibid., p. 47.

[5] Immanuel Kant, “On a Supposed Right to Lie Because of Philanthropic Concerns,” Berlinische Blätter, 1799.

[6] Larry May, War Crimes and Just War, Cambridge University Press, 2007, pp. 196-197.