Arguments for Utilitarianism

Introduction: Moral Methodology & Reflective Equilibrium

You cannot prove a moral theory. Whatever arguments you come up with, it’s always possible for someone else to reject your premises—if they are willing to accept the costs of doing so. Different theories offer different advantages. This chapter will set out some of the major considerations that plausibly count in favor of utilitarianism. A complete view also needs to consider the costs of utilitarianism (or the advantages of its competitors), which are addressed in Chapter 8: Objections to Utilitarianism. You can then reach an all-things-considered judgment as to which moral theory strikes you as overall best or most plausible.

To this end, moral philosophers typically use the methodology of reflective equilibrium.¹ This involves balancing two broad kinds of evidence as applied to moral theories:

1. Intuitions about specific cases (thought experiments).
2. General theoretical considerations, including the plausibility of the theory’s principles or systematic claims about what matters.

General principles can be challenged by coming up with putative counterexamples, or cases in which they give an intuitively incorrect verdict. In response to such putative counterexamples, we must weigh the force of the case-based intuition against the inherent plausibility of the principle being challenged. This could lead you to either revise the principle to accommodate your intuitions about cases or to reconsider your verdict about the specific case, if you judge the general principle to be better supported (especially if you are able to “explain away” the opposing intuition as resting on some implicit mistake or confusion).

As we will see, the arguments in favor of utilitarianism rest overwhelmingly on general theoretical considerations. Challenges to the view can take either form, but many of the most pressing objections involve thought experiments in which utilitarianism is held to yield counterintuitive verdicts.

There is no neutral, non-question-begging answer to how one ought to resolve such conflicts.² It takes judgment, and different people may be disposed to react in different ways depending on their philosophical temperament. As a general rule, those of a temperament that favors systematic theorizing are more likely to be drawn to utilitarianism (and related views), whereas those who hew close to common sense intuitions are less likely to be swayed by its theoretical virtues. Considering the arguments below may thus do more than just illuminate utilitarianism; it may also help you to discern your own philosophical temperament!

While our presentation focuses on utilitarianism, it’s worth noting that many of the arguments below could also be taken to support other forms of welfarist consequentialism (just as many of the objections to utilitarianism also apply to these related views). This chapter explores arguments for utilitarianism and closely related views over non-consequentialist approaches to ethics.

Arguments for Utilitarianism

What Fundamentally Matters

Moral theories serve to specify what fundamentally matters, and utilitarianism offers a particularly compelling answer to this question.

Almost anyone would agree with utilitarianism that suffering is bad, and well-being is good. What could be more obvious? If anything matters morally, human well-being surely does. And it would be arbitrary to limit moral concern to our own species, so we should instead conclude that well-being generally is what matters. That is, we ought to want the lives of sentient beings to go as well as possible (whether that ultimately comes down to maximizing happiness, desire satisfaction, or other welfare goods).

Could anything else be more important? Such a suggestion can seem puzzling. Consider: it is (usually) wrong to steal.³ But that is plausibly because stealing tends to be harmful, reducing people’s well-being.⁴ By contrast, most people are open to redistributive taxation, if it allows governments to provide benefits that reliably raise the overall level of well-being in society. So it’s not that individuals just have a natural right to not be interfered with no matter what. When judging institutional arrangements (such as property and tax law), we recognize that what matters is coming up with arrangements that tend to secure overall good results, and that the most important factor in what makes a result good is that it promotes well-being.⁵

Such reasoning may justify viewing utilitarianism as the default starting point for moral theorizing.⁶ If someone wants to claim that there is some other moral consideration that can override overall well-being (trumping the importance of saving lives, reducing suffering, and promoting flourishing), they face the challenge of explaining how that could possibly be so. Many common moral rules (like those that prohibit theft, lying, or breaking promises), while not explicitly utilitarian in content, nonetheless have a clear utilitarian rationale. If they did not generally promote well-being—but instead actively harmed people—it’s hard to see what reason we would have to still want people to follow them. To follow and enforce harmful moral rules (such as rules prohibiting same-sex relationships) would seem like a kind of “rule worship”, and not truly ethical at all.⁷ Since the only moral rules that seem plausible are those that tend to promote well-being, that’s some reason to think that moral rules are, as utilitarianism suggests, purely instrumental to promoting well-being.

Similar judgments apply to hypothetical cases in which you somehow know for sure that a typically reliable rule is, in this particular instance, counterproductive. In the extreme case, we all recognize that you ought to lie or break a promise if lives are on the line. In practice, of course, the best way to achieve good results over the long run is to respect commonsense moral rules and virtues while seeking opportunities to help others. (It’s important not to mistake the hypothetical verdicts utilitarianism offers in stylized thought experiments with the practical guidance it offers in real life.) The key point is just that utilitarianism offers a seemingly unbeatable answer to the question of what fundamentally matters: protecting and promoting the interests of all sentient beings to make the world as good as it can be.

The Veil of Ignorance

Humans are masters of self-deception and motivated reasoning. If something benefits us personally, it’s all too easy to convince ourselves that it must be okay. We are also more easily swayed by the interests of more salient or sympathetic individuals (favoring puppies over pigs, for example). To correct for such biases, it can be helpful to force impartiality by imagining that you are looking down on the world from behind a “veil of ignorance”. This veil reveals the facts about each individual’s circumstances in society—their income, happiness level, preferences, etc.—and the effects that each choice would have on each person, while hiding from you the knowledge of which of these individuals you are.⁸ To more fairly determine what ideally ought to be done, we may ask what everyone would have most personal reason to prefer from behind this veil of ignorance. If you’re equally likely to end up being anyone in the world, it would seem prudent to maximize overall well-being, just as utilitarianism prescribes.⁹

It’s an interesting question how much weight we should give to the verdicts that would be chosen, on self-interested grounds, from behind the veil. The veil thought experiment serves to highlight how utilitarianism gives equal weight to everyone’s interests, in unbiased fashion. That is, utilitarianism is just what we get when we are beneficent to all: extending to everyone the kind of careful concern that prudent people have for their own interests.¹⁰ But it may seem question-begging to those who reject welfarism, and so deny that interests are all that matter. For example, the veil thought experiment clearly doesn’t speak to the question of whether non-sentient life or natural beauty has intrinsic value. It’s restricted to that sub-domain of morality that concerns what we owe to each other, where this includes just those individuals over whom our veil-induced uncertainty about our identity extends: presently existing sentient beings, perhaps.¹¹ Accordingly, any verdicts reached on the basis of the veil of ignorance will still need to be weighed against what we might yet owe to any excluded others (such as future generations, or non-welfarist values).

Still, in many contexts other factors will not be relevant, and the question of what we morally ought to do will reduce to the question of how we should treat each other. Many of the deepest disagreements between utilitarians and their critics concern precisely this question. And the veil of ignorance seems relevant here. The fact that some action is what everyone affected would personally prefer from behind the veil of ignorance seems to undermine critics’ claims that any individual has been mistreated by, or has grounds to complain about, that action.

Ex Ante Pareto

A Pareto improvement is better for some people, and worse for none. When outcomes are uncertain, we may instead assess the prospect associated with an action—the range of possible outcomes, weighted by their probabilities. A prospect can be assessed as better for you when it offers you greater well-being in expectation, or ex ante.¹² Putting these concepts together, we may formulate the following principle:

Ex ante Pareto: in a choice between two prospects, one is morally preferable to another if it offers a better prospect for some individuals and a worse prospect for none.

This bridge between personal value (or well-being) and moral assessment is further developed in economist John Harsanyi’s aggregation theorem.¹³ But the underlying idea, that reasonable beneficence requires us to wish well to all, and prefer prospects that are in everyone’s ex ante interests, has also been defended and developed in more intuitive terms by philosophers.¹⁴

A powerful objection to most non-utilitarian views is that they sometimes violate ex ante Pareto, such as when choosing policies from behind the veil of ignorance. Many rival views imply, absurdly, that prospect Y could be morally preferable to prospect X, even when Y is worse in expectation for everyone involved.

Caspar Hare illustrates the point with a Trolley case in which all six possible victims are stuffed inside suitcases: one is atop a footbridge, five are on the tracks below, and a train will hit and kill the five unless you topple the one on the footbridge (in which case the train will instead kill this one and then stop before reaching the others).¹⁵ As the suitcases have recently been shuffled, nobody knows which position they are in. So, from each victim’s perspective, their prospects are best if you topple the one suitcase off the footbridge, increasing their chances of survival from 1/6 to 5/6. Given that this is in everyone’s ex ante interests, it’s deeply puzzling to think that it would be morally preferable to override this unanimous preference, shared by everyone involved, and instead let five of the six die; yet that is the implication of most non-utilitarian views.¹⁶

Expanding the Moral Circle

When we look back on past moral atrocities—like slavery or denying women equal rights—we recognize that they were often sanctioned by the dominant societal norms at the time. The perpetrators of these atrocities were grievously wrong to exclude their victims from their “circle” of moral concern.¹⁷ That is, they were wrong to be indifferent towards (or even delight in) their victims’ suffering. But such exclusion seemed normal to people at the time. So we should question whether we might likewise be blindly accepting of some practices that future generations will see as evil but that seem “normal” to us.¹⁸ The best protection against making such an error ourselves would be to deliberately expand our moral concern outward, to include all sentient beings—anyone who can suffer—and so recognize that we have strong moral reasons to reduce suffering and promote well-being wherever we can, no matter who it is that is experiencing it.

While this conclusion is not yet all the way to full-blown utilitarianism, since it’s compatible with, for example, holding that there are side-constraints limiting one’s pursuit of the good, it is likely sufficient to secure agreement with the most important practical implications of utilitarianism (stemming from cosmopolitanism, anti-speciesism, and longtermism).

The Poverty of the Alternatives

We’ve seen that there is a strong presumptive case in favor of utilitarianism. If no competing view can be shown to be superior, then utilitarianism has a strong claim to be the “default” moral theory. In fact, one of the strongest considerations in favor of utilitarianism (and related consequentialist views) is the deficiencies of the alternatives. Deontological (or rule-based) theories, in particular, seem to rest on questionable foundations.¹⁹

Deontological theories are explicitly non-consequentialist: instead of morally assessing actions by evaluating their consequences, these theories tend to take certain types of action (such as killing an innocent person) to be intrinsically wrong.²⁰ There are reasons to be dubious of this approach to ethics, however.

The Paradox of Deontology

Deontologists hold that there is a constraint against killing: that it’s wrong to kill an innocent person even if this would save five other innocent people from being killed. This verdict can seem puzzling on its face.²¹ After all, given how terrible killing is, should we not want there to be less of it? Rational choice in general tends to be goal-directed, a conception which fits poorly with deontic constraints.²² A deontologist might claim that their goal is simply to avoid violating moral constraints themselves, which they can best achieve by not killing anyone, even if that results in more individuals being killed. While this explanation can render deontological verdicts coherent, it does so at the cost of making them seem awfully narcissistic, as though the deontologist’s central concern was just to maintain their own moral purity or “clean hands”.

Deontologists might push back against this characterization by instead insisting that moral action need not be goal-directed at all.²³ Rather than only seeking to promote value (or minimize harm), they claim that moral agents may sometimes be called upon to respect another’s value (by not harming them, even as a means to preventing greater harm to others), which would seem an appropriately outwardly-directed, non-narcissistic motivation.

The challenge remains that such a proposal makes moral norms puzzlingly divergent from other kinds of practical norms. If morality sometimes calls for respecting value rather than promoting it, why is the same not true of prudence? (Given that pain is bad for you, for example, it would not seem prudent to refuse a painful operation now if the refusal commits you to five comparably painful operations in future.) Deontologists may offer various answers to this question, but insofar as we are inclined to think, pre-theoretically, that ethics ought to be continuous with other forms of rational choice, that gives us some reason to prefer consequentialist accounts.

Deontologists also face a tricky question about where to draw the line. Is it at least okay to kill one person to prevent a hundred killings? Or a million? Absolutists never permit killing, no matter the stakes. But such a view seems too extreme for many. Moderate deontologists allow that sufficiently high stakes can justify violations. But how high? Any answer they offer is apt to seem arbitrary and unprincipled. Between the principled options of consequentialism or absolutism, many will find consequentialism to be the more plausible of the two.

The Hope Objection

Impartial observers should want and hope for the best outcome. Non-consequentialists claim that nonetheless it’s sometimes wrong to bring about the best outcome. Putting the two claims together yields the striking result that you should sometimes hope that others act wrongly.

Suppose it would be wrong for some stranger—call him Jack—to kill one innocent person to prevent five other (morally comparable) killings. Non-consequentialists may claim that Jack has a special responsibility to ensure that he does not kill anyone, even if this results in more killings by others. But you are not Jack. From your perspective as an impartial observer, Jack’s killing one innocent person is no more or less intrinsically bad than any of the five other killings that would thereby be prevented. You have most reason to hope that there is only one killing rather than five. So you have reason to hope that Jack acts “wrongly” (killing one to save five). But that seems odd.

More than merely being odd, this might even be taken to undermine the claim that deontic constraints matter, or are genuinely important to abide by. After all, to be important just is to be worth caring about. For example, we should care if others are harmed, which validates the claim that others’ interests are morally important. But if we should not care more about Jack’s abiding by the moral constraint against killing than we should about his saving five lives, that would seem to suggest that the constraint against killing is not in fact more morally important than saving five lives.

Finally, since our moral obligations ought to track what is genuinely morally important, if deontic constraints are not in fact important then we cannot be obligated to abide by them.²⁴ We cannot be obliged to prioritize deontic constraints over others’ lives, if we ought to care more about others’ lives than about deontic constraints. So deontic constraints must not accurately describe our obligations after all. Jack really ought to do whatever would do the most good overall, and so should we.

Skepticism About the Distinction Between Doing and Allowing

You might wonder: if respect for others requires not harming them (even to help others more), why does it not equally require not allowing them to be harmed? Deontological moral theories place great weight on distinctions such as those between doing and allowing harm, or killing and letting die, or intended versus merely foreseen harms. But why should these be treated so differently? If a victim ends up equally dead either way, whether they were killed or “merely” allowed to die would not seem to make much difference to them—surely what matters to them is just their death. Consequentialism accordingly denies any fundamental significance to these distinctions.²⁵

Indeed, it’s far from clear that there is any robust distinction between “doing” and “allowing”. Sometimes you might “do” something by remaining perfectly still.²⁶ Also, when a doctor unplugs a terminal patient from life support machines, this is typically thought of as “letting die”; but if a mafioso, worried about an informant’s potentially incriminating testimony, snuck in to the hospital and unplugged the informant’s life support, we are more likely to judge it to constitute “killing”.²⁷ Bennett (1998) argues at length that there is no satisfactory, fully general distinction between doing and allowing—at least, none that would vindicate the moral significance that deontologists want to attribute to such a distinction.²⁸ If Bennett is right, then that might force us towards some form of consequentialism (such as utilitarianism) instead.

Status Quo Bias

Opposition to utilitarian trade-offs—that is, benefiting some at a lesser cost to others—arguably amounts to a kind of status quo bias, prioritizing the preservation of privilege over promoting well-being more generally.

Such conservatism might stem from the Just World fallacy: the mistake of assuming that the status quo is just, and that people naturally get what they deserve. Of course, reality offers no such guarantees of justice. What circumstances one is born into depends on sheer luck, including one’s endowment of physical and cognitive abilities which may pave the way for future success or failure. Thus, even later in life we never manage to fully wrest back control from the whimsies of fortune and, consequently, some people are vastly better off than others despite being no more deserving. In such cases, why should we not be willing to benefit one person at a lesser cost to privileged others? They have no special entitlement to the extra well-being that fortune has granted them.²⁹ Clearly, it’s good for people to be well-off, and we certainly would not want to harm anyone unnecessarily.³⁰ However, if we can increase overall well-being by benefiting one person at the lesser cost to another, we should not refrain from doing so merely due to a prejudice in favor of the existing distribution.³¹ It’s easy to see why traditional elites would want to promote a “morality” which favors their entrenched interests. It’s less clear why others should go along with such a distorted view of what (and who) matters.

It can similarly be argued that there is no real distinction between imposing harms and withholding benefits. The only difference between the two cases concerns what we understand to be the status quo, which lacks moral significance. Suppose scenario A is better for someone than B. Then to shift from A to B would be a “harm”, while to prevent a shift from B to A would be to “withhold a benefit”. But this is merely a descriptive difference. If we deny that the historically given starting point provides a morally privileged baseline, then we must say that the cost in either case is the same, namely the difference in well-being between A and B. In principle, it should not matter where we start from.³²

Now suppose that scenario B is vastly better for someone else than A is: perhaps it will save their life, at the cost of the first person’s arm. Nobody would think it okay to kill a person just to save another’s arm (that is, to shift from B to A). So if we are to avoid status quo bias, we must similarly judge that it would be wrong to oppose the shift from A to B—that is, we should not object to saving someone’s life at the cost of another’s arm.³³ We should not care especially about preserving the privilege of whoever stood to benefit by default; such conservatism is not truly fair or just. Instead, our goal should be to bring about whatever outcome would be best overall, counting everyone equally, just as utilitarianism prescribes.

Evolutionary Debunking Arguments

Against these powerful theoretical objections, the main consideration that deontological theories have going for them is closer conformity with our intuitions about particular cases. But if these intuitions cannot be supported by independently plausible principles, that may undermine their force—or suggest that we should interpret these intuitions as good rules of thumb for practical guidance, rather than as indicating what fundamentally matters.

The force of deontological intuitions may also be undermined if it can be demonstrated that they result from an unreliable process. For example, evolutionary processes may have endowed us with an emotional bias favoring those who look, speak, and behave like ourselves; this, however, offers no justification for discriminating against those unlike ourselves. Evolution is a blind, amoral process whose only “goal” is the propagation of genes, not the promotion of well-being or moral rightness. Our moral intuitions require scrutiny, especially in scenarios very different from our evolutionary environment. If we identify a moral intuition as stemming from our evolutionary ancestry, we may decide not to give much weight to it in our moral reasoning—the practice of evolutionary debunking.³⁴

Katarzyna de Lazari-Radek and Peter Singer argue that views permitting partiality are especially susceptible to evolutionary debunking, whereas impartial views like utilitarianism are more likely to result from undistorted reasoning.³⁵ Joshua Greene offers a different psychological debunking argument. He argues that deontological judgments—for instance, in response to trolley cases—tend to stem from unreliable and inconsistent emotional responses, including our favoritism of identifiable over faceless victims and our aversion to harming someone up close rather than from afar. By contrast, utilitarian judgments involve the more deliberate application of widely respected moral principles.³⁶

Such debunking arguments raise worries about whether they “prove too much”: after all, the foundational moral judgment that pain is bad would itself seem emotionally-laden and susceptible to evolutionary explanation—physically vulnerable creatures would have powerful evolutionary reasons to want to avoid pain whether or not it was objectively bad, after all!³⁷

However, debunking arguments may be most applicable in cases where we feel that a principled explanation for the truth of the judgment is lacking. We do not tend to feel any such lack regarding the badness of pain—that is surely an intrinsically plausible judgment if anything is. Some intuitions may be over-determined: explicable both by evolutionary causes and by their rational merits. In such a case, we need not take the evolutionary explanation to undermine the judgment, because the judgment also results from a reliable process (namely, rationality). By contrast, deontological principles and partiality are far less self-evidently justified, and so may be considered more vulnerable to debunking. Once we have an explanation for these psychological intuitions that can explain why we would have them even if they were rationally baseless, we may be more justified in concluding that they are indeed rationally baseless.

As such, debunking objections are unlikely to change the mind of one who is drawn to the target view (or regards it as independently justified and defensible). But they may help to confirm the doubts of those who already felt there were some grounds for scepticism regarding the intrinsic merits of the target view.

Conclusion

Utilitarianism can be supported by several theoretical arguments, the strongest perhaps being its ability to capture what fundamentally matters. Its main competitors, by contrast, seem to rely on dubious distinctions—like “doing” vs. “allowing”—and built-in status quo bias. At least, that is how things are apt to look to one who is broadly sympathetic to a utilitarian approach. Given the flexibility inherent in reflective equilibrium, these arguments are unlikely to sway a committed opponent of the view. For those readers who find a utilitarian approach to ethics deeply unappealing, we hope that this chapter may at least help you to better understand what appeal others might see in the view.

However strong you judge the arguments in favor of utilitarianism to be, your ultimate verdict on the theory will also depend upon how well the view is able to counter the influential objections that critics have raised against it.

The next chapter discusses theories of well-being, or what counts as being good for an individual.

Resources and Further Reading

• John Broome (1987). Utilitarianism and Expected Utility, The Journal of Philosophy 84 (8): 405–422.
• John Broome (1991). Weighing Goods: Equality, Uncertainty and Time. Blackwell.
• Krister Bykvist (2010). Utilitarianism: A Guide for the Perplexed. Continuum.
• Robert Goodin (1995). Utilitarianism as a Public Philosophy. Cambridge University Press.
• Johan Gustafsson (2021). Utilitarianism without Moral Aggregation. Canadian Journal of Philosophy 51 (4): 256-269.
• Caspar Hare (2016). Should We Wish Well to All?, Philosophical Review 125(4): 451–472.
• John C. Harsanyi (1955). Cardinal Welfare, Individualistic Ethics, and Interpersonal Comparisons of Utility, The Journal of Political Economy 63 (4): 309–321.
• John C. Harsanyi (1977). Rational Behavior and Bargaining Equilibrium in Games and Social Situations. Cambridge University Press.
• Katarzyna de Lazari-Radek & Peter Singer (2017). Chapter 2: Justifications, in Utilitarianism: A Very Short Introduction. Oxford University Press.
• J.J.C. Smart (1973). An outline of a system of utilitarian ethics, in J.J.C. Smart & Bernard Williams, Utilitarianism: For and Against. Cambridge University Press.