Ethics: Consequentialism, Deontology, and Virtue

What It Is

Normative ethics is the study of what makes actions right or wrong, what kinds of lives are good or bad, and what counts as a good character. Contemporary normative ethics is dominated by three large families of theory. Each family proposes a different kind of fact as the right-maker: outcomes, duty-conforming features of the act, or features of the agent's character.

• Consequentialism. An act is right iff it produces the best outcome among the available alternatives. The most prominent variant is utilitarianism: the best outcome is the one with the greatest aggregate well-being. Other variants substitute different value-bearers (preference satisfaction, capability, objective goods).
• Deontology. An act is right iff it conforms to certain duty-defining features, irrespective of outcomes (or with only limited weight to outcomes). The most prominent variant is Kantian ethics: the right act is the one whose maxim could be willed as a universal law and that treats persons always as ends in themselves, never merely as means.
• Virtue ethics. An act is right iff it is what a fully virtuous agent would characteristically do in the circumstances. The primary subject of ethical evaluation is character (the stable dispositions of a person), not act. The most prominent variant traces to Aristotle's Nicomachean Ethics.

These are structurally distinct theories, not just different labels for similar conclusions. They identify different facts as right-makers, they generate predictably different verdicts on stock cases, and they imply different patterns of moral reasoning. The contemporary debate is largely about which family captures what we already implicitly believe, and how the families can be combined or refined.

This page assumes What Is Philosophy?. The intersection with Free Will and Determinism is the question of moral responsibility (responsibility presupposes some account of agency); the intersection with Personal Identity Over Time is the question of who bears responsibility across time.

Consequentialism

The clearest statement is in Jeremy Bentham's Introduction to the Principles of Morals and Legislation (1789): the principle of utility approves or disapproves every action according to its tendency to augment or diminish the happiness of those whose interest is in question. J.S. Mill refined the doctrine in Utilitarianism (1861): the right action is the one that maximizes happiness, where happiness is "pleasure and the absence of pain"; Mill admitted qualitative distinctions among pleasures.¹

The general schema. Pick a value theory (what makes outcomes good: hedonic utility, preference satisfaction, well-being, capabilities, objective goods). Pick a target (maximize total, average, weighted, sufficientarian, prioritarian). The right act maximizes the chosen value-aggregate.

Act consequentialism. The right act is the one whose actual or expected outcome maximizes the value-aggregate. Decision: at each choice point, perform the act with the best expected outcome.

Rule consequentialism. The right rule is the rule whose general adoption would maximize the value-aggregate; the right act is the act required by that rule. Decision: at each choice point, follow the optimal rule, even when departing from it would produce a better outcome in this case.

Three serious objections to consequentialism.

Demandingness. Consequentialism appears to demand that the agent maximize the good in every act, with no special weight for the agent's own projects or relationships. The agent must always do the most good available, even if that requires sacrificing personal interests. Bernard Williams argued (1973²) that this is too strong: a coherent ethical agent must have integrity-grounding personal projects, and a theory that demands total impartiality is psychologically untenable.

The separateness of persons. Consequentialism treats well-being as a single aggregate to be maximized. But persons are separate: harming person $A$ to benefit person $B$ does not literally net out in the way harming $A$ in one moment to benefit $A$ in another might. Rawls (1971³) argued that consequentialism "does not take seriously the distinction between persons"; the same theoretical structure used for intra-personal trade-offs (where it is licensed) is used for inter-personal trade-offs (where it is not).

The repugnant conclusion (Parfit 1984). A pure total utilitarianism implies that a world with vastly many people each at a barely-worth-living quality of life is better than a smaller world with everyone flourishing, provided the total sum is higher. Parfit called this "the repugnant conclusion" and treated it as a reductio of any pure total view. The literature on population ethics has been wrestling with the conclusion ever since.

Consequentialist responses include: (i) refining the value theory (objective-list theories, prioritarianism that weights worse-off persons more heavily); (ii) refining the target (averagism, sufficientarianism); (iii) accepting the conclusion as counterintuitive but not refuting; (iv) moving to rule consequentialism, which generates much weaker demandingness and integrity worries.

Deontology

Immanuel Kant's Groundwork of the Metaphysics of Morals (1785) is the canonical reference. Kant argued that the moral worth of an action is fixed by the maxim (the subjective principle) on which the agent acts, not by its consequences. A moral maxim must be universalizable: an agent can rationally act on a maxim iff she can will that the maxim become a universal law of nature without contradiction.

The Categorical Imperative has three main formulations.

• Universal law formulation. Act only on that maxim through which you can at the same time will that it should become a universal law.
• Humanity formulation. Act in such a way that you treat humanity, whether in your own person or in the person of another, always at the same time as an end and never merely as a means.
• Kingdom of ends formulation. Act as if you were always through your maxims a law-making member in a universal kingdom of ends.

The three formulations are intended by Kant as alternative routes to the same content, though their equivalence is contested in the literature.

Modern deontological theories take different shapes.

Threshold deontology (Thomson, Nagel). Duty-defined constraints hold against producing the best outcome, but the constraints are not absolute. At some sufficiently catastrophic threshold, the constraint can be overridden.

Rights-based deontology (Nozick 1974⁴). Persons have rights that function as side constraints on what others may do to them. The right not to be killed for organ harvesting is a side constraint even if the killing would save five lives by harvest.

Scanlonian contractualism (Scanlon 1998⁵). An act is wrong iff its performance would be disallowed by any principle for the general regulation of behavior that no one could reasonably reject. Contractualism shares the deontological spirit (rightness is not fixed by outcomes) while grounding the constraints in justifiability-to-each-affected-person.

Three serious objections to deontology.

Arbitrariness. Why these duties and not others? Kantians appeal to universalizability, but the universalizability test seems to license different verdicts depending on how the maxim is formulated. Rights-based theorists must explain where the rights come from. Without a deeper foundation the duty-set looks arbitrary.

Catastrophic outcomes. Kant famously argued one must not lie to the murderer at the door asking whether the friend they wish to kill is in the house. Strict deontology produces verdicts most agents find unacceptable in extreme cases. The threshold-deontology response builds in an outcome-sensitive escape clause but then owes an account of where the threshold lies.

The trolley distinction. Most deontologists treat diverting a trolley (foreseeably killing one to save five) as permissible but pushing the large person off the bridge (intentionally killing one to save five) as impermissible. The doctrine of double effect (DDE) is invoked: intending harm as a means is impermissible, foreseeing harm as a side effect is permissible. The distinction is contested: critics argue DDE-style distinctions are not as principled as they look and can be redescribed to flip the verdict.

Virtue Ethics

Aristotle's Nicomachean Ethics is the canonical source. The central thesis: the primary subject of ethical evaluation is character (the stable dispositions of a person), and the right action is the action that the virtuous agent would characteristically perform. The virtues are stable dispositions to feel, act, and choose in the right way relative to a kind of situation; they are acquired by habituation and refined by phronesis (practical wisdom).⁶

Five features distinguish virtue ethics from consequentialism and deontology.

1. Character primary. Ethical evaluation is fundamentally of persons, not of acts. Acts are evaluated through their connection to character (an act is courageous because a courageous person would do it).
2. Agent-relative. The right answer for me depends partly on what kind of agent I am. There is no impartial agent-neutral verdict.
3. Particularist. The right action is sensitive to the particular circumstances; there are no useful general action-prescribing rules in the way that Kantian universalizability or utility-maximization purports to provide.
4. Eudaimonist. The framework is grounded in eudaimonia (flourishing, the good life). The virtues are good because they are constitutive of (or instrumentally necessary for) the flourishing life.
5. Habituation-based. Virtues are acquired by practice, not by inferring from first principles. Moral education is moral training.

Contemporary revival starts with G.E.M. Anscombe's "Modern Moral Philosophy" (1958⁷), which argued that the modern act-evaluation frameworks rest on a confused notion of moral obligation whose original theological grounding has been lost. Alasdair MacIntyre's After Virtue (1981) extended the critique. Rosalind Hursthouse's On Virtue Ethics (1999) provides the canonical contemporary statement.

Three serious objections to virtue ethics.

Action guidance. The framework tells you to do what the virtuous agent would do. But what does the virtuous agent do? The view risks being uninformative: it tells the deliberating agent to consult her own moral exemplars or her cultivated phronesis, but does not give an external criterion against which to check the answer.

Circularity. A virtuous person is one who does the right thing; the right thing is what the virtuous person does. Without an independent criterion for either, the theory is circular. Hursthouse's response (1999) is to ground the virtues in a substantive theory of human flourishing; whether this avoids the circularity is debated.

The empirical challenge (Doris 1998, Harman 1999). Empirical psychology suggests that situations explain more variance in behavior than stable character traits do. The classic Hartshorne-May studies of moral character in children, the Milgram obedience experiments, and the Stanford prison experiment all suggest that what philosophers call "character" is much less stable than virtue ethics supposes. If virtues are not the empirically real things virtue ethics needs them to be, the framework loses traction.

Virtue-ethicist responses include reinterpreting the empirical evidence (situations matter, but so does character at coarser timescales), refining the conception of virtue (virtues are local and context-bound, not global), or arguing that the empirical critique trades on a thin conception of virtue.

Worked Case: The Trolley

A runaway trolley is heading down a track toward five workers who will be killed unless something is done. You can pull a switch that diverts the trolley to a side track, where it will kill one worker but save the five.

• Consequentialist verdict. Pull the switch. Five lives saved versus one lost; the outcome is straightforwardly better. (Some refinements: rule consequentialism might generate the same verdict via a rule of action; sophisticated consequentialism may add agent-relative considerations.)
• Deontological verdict (standard). Pull the switch is permissible but not required. The death of the one is a foreseen but not intended side effect of an action whose goal is saving the five. Doctrine of double effect licenses the act.
• Virtue ethics verdict. What would a phronetic agent do, given courage, justice, compassion, and the actual circumstances? The view does not give a clean derivation; it asks what the well-developed moral perception of the agent recommends in the particular case. Most virtue ethicists agree the switch should be pulled but emphasize the difficulty of the case and the importance of the agent's grief and ongoing concern for the family of the one killed.

Now consider the footbridge variant. You are standing on a footbridge over the trolley track. Beside you is a large person whose body, pushed off the bridge, would stop the trolley and save the five (you are too small to stop it yourself).

• Consequentialist verdict. Push. Same arithmetic.
• Deontological verdict. Do not push. The death of the large person is intended as a means to saving the five, not merely foreseen as a side effect. The doctrine of double effect distinguishes the cases.
• Virtue ethics verdict. A phronetic agent recognizes pushing as a different kind of action from diverting; the difference is what the action of pushing expresses and what it shows about the agent. Most virtue ethicists conclude do not push.

The three families pull apart on the footbridge case but not on the original. The pulling-apart is what makes the case philosophically informative: it isolates the consequentialist's structural commitment to outcome-only evaluation from the deontologist's structural commitment to means / side effect distinctions and from virtue ethics's structural commitment to act-as-expression-of-character.

Common Misconceptions

• "Utilitarianism is the same as consequentialism." No. Utilitarianism is one consequentialist theory, distinguished by its commitment to a hedonic or welfare-based value theory and a maximizing aggregation rule. Consequentialism is the broader genus that also includes theories with different value theories (objective-list, preference-satisfaction) and different aggregation rules (prioritarian, sufficientarian).
• "Deontology means following rules without thinking." No. Kantian ethics is a theory about what rationality requires of an agent in choice; the categorical imperative is the form of the test, not a list of rules. Sophisticated deontologists have refined views about how rules apply in particular cases, how exceptions are handled, and how conflicting duties are resolved.
• "Virtue ethics is conservative because it valorizes traditional virtues." Not necessarily. The contemporary literature has expanded the catalog of virtues (intellectual virtues, civic virtues, virtues specific to professional roles) and has been used to argue for revisionary positions (animal ethics, environmental ethics). The framework's commitment to character and habituation is structural; the particular catalog of virtues is empirical.
• "The three families are just different decision procedures for the same conclusions." No. They are structurally different theories: they identify different facts as right-makers, they generate different verdicts on cases like footbridge, and they imply different patterns of moral reasoning. Treating them as procedurally equivalent erases what the disagreement is about.

Comparisons to Related Views

Go Further

• Stanford Encyclopedia of Philosophy, "Consequentialism" by Walter Sinnott-Armstrong. The clearest survey of the family.
• Stanford Encyclopedia of Philosophy, "Deontological Ethics" by Larry Alexander and Michael Moore.
• Stanford Encyclopedia of Philosophy, "Virtue Ethics" by Rosalind Hursthouse and Glen Pettigrove. Hursthouse is the author of the canonical contemporary statement.
• Parfit, Derek. On What Matters. 3 vols. Oxford University Press, 2011–2017. Parfit's late attempt to argue the three traditions converge; widely influential, widely contested.
• Hursthouse, Rosalind. On Virtue Ethics. Oxford University Press, 1999. The reference virtue-ethics text.
• Scanlon, T.M. What We Owe to Each Other. Harvard University Press, 1998. The reference contractualist text.
• Driver, Julia. Ethics: The Fundamentals. Blackwell, 2006. The cleanest single textbook introduction.

For the moral-responsibility presupposition behind any normative ethics, see Free Will and Determinism. For the question of who bears responsibility across time, see Personal Identity Over Time.