{
  "id": "consequentialism",
  "title": "Consequentialism",
  "lang": "en",
  "source_lang": "en",
  "summary": "Consequentialism holds that moral rightness depends on consequences, with competing theories specifying which outcomes matter and how they should be assessed.",
  "category": "philosophy",
  "subcategory": "ethics",
  "type": "concept",
  "aliases": [
    "Consequentialist ethics"
  ],
  "generated_at": "2026-11-04",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/consequentialism/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/consequentialism/",
      "json": "https://aiwiki.page/zh/consequentialism.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "Consequentialism is a family of theories in ethics according to which the moral rightness of actions, rules, or other objects of evaluation depends entirely on their consequences. Its best-known form, utilitarianism, judges conduct by its contribution to overall well-being. Consequentialism is broader, however: a theory may evaluate outcomes in terms of equality, knowledge, freedom, or several goods together rather than happiness alone.\n\nConsequentialist theories differ over what counts as a good outcome, whose interests count, whether agents must maximize goodness, and whether individual actions or general rules should be assessed. They also distinguish judgments based on actual results from judgments based on reasonable expectations. These differences affect how the theories address uncertainty, personal relationships, rights, and obligations to future generations.\n\nConsequentialism is commonly contrasted with deontology, which recognizes duties or constraints not wholly grounded in outcome value, and with virtue ethics, which gives a central place to character and human flourishing. These boundaries are not absolute: consequentialists can defend rules and virtues, while nonconsequentialists can regard consequences as morally important. The distinctive question is whether consequences provide the ultimate basis of moral justification."
    },
    {
      "heading": "The structure of consequentialist justification",
      "text": "A consequentialist theory combines an account of value with a principle connecting value to right action. Its evaluative component specifies which outcomes are better or worse; its normative component determines what an agent ought to do given that ranking. In a simple maximizing formulation, an action is right if no available alternative has better consequences. If several alternatives are equally best, each may be permissible. Consequently, maximizing consequentialism need not prescribe a uniquely correct action in every situation.\n\nThe structure can be represented formally. Let \\(A\\) be the set of actions available to an agent, let \\(C(a)\\) represent the consequences of action \\(a\\), and let \\(V\\) evaluate those consequences. Then:\n\n\\[a \\text{ is right if and only if } V(C(a))=\\max_{b\\in A}V(C(b)).\\]\n\nThis expression does not yet identify goodness with pleasure, require monetary measurement, or specify how uncertainty should be handled. It also leaves open how the relevant alternatives are defined. An action might be physically possible but unavailable because the agent lacks the necessary information, authority, or opportunity. Different specifications can change the theory’s verdict.\n\nConsequences are not limited to immediate or intended effects. They may include long-term changes in trust, incentives, institutions, expectations, and character. Nevertheless, consequentialism is not the claim that any successful action is morally justified. Achieving an agent’s private objective is different from producing a morally best outcome, and the relevant comparison concerns available alternatives rather than the desirability of one result considered alone. Explaining what would happen under each alternative requires judgments about causation and counterfactual dependence, not simply a list of events that followed an action.\n\nStandard utilitarian theories evaluate outcomes impartially: a benefit does not count more simply because its recipient is the person acting. Some broader consequentialist theories allow agent-relative rankings, under which different agents may appropriately give different weights to particular relationships or responsibilities. The classification of such theories is disputed. If any duty can be represented as a sufficiently strong preference over outcomes, consequentialism risks becoming a formal restatement of other moral views rather than a distinctive theory. Much therefore depends on whether outcome value can be specified independently of the obligations it is meant to explain."
    },
    {
      "heading": "Historical development and terminology",
      "text": "Arguments resembling consequentialist reasoning appear in several philosophical traditions, but it is important to distinguish attention to beneficial results from a complete consequentialist theory. Ancient Chinese Mohism assessed practices partly by their contribution to order, material provision, and population, and opposed costly warfare and elaborate funerals on grounds of public benefit. Its doctrines of impartial concern and practical benefit provide significant parallels with later consequentialist thought. They nevertheless developed within a framework that also included Heaven, spirits, and political authority, rather than the conceptual setting of modern European moral philosophy.\n\nModern utilitarianism emerged from eighteenth-century debates about benevolence, pleasure, legislation, and social improvement. David Hume emphasized the usefulness of qualities and institutions in explaining moral approval, although his moral philosophy cannot simply be identified with maximizing consequentialism. Jeremy Bentham’s An Introduction to the Principles of Morals and Legislation, published in 1789, presented the principle of utility as a basis for judging actions and governmental measures. His discussion of pleasure and pain considered features including intensity, duration, certainty, and extent—the last concerning how many individuals are affected.\n\nJohn Stuart Mill developed utilitarianism in essays first published in 1861 and subsequently issued as a book in 1863. Mill defended qualitative differences among pleasures and connected utility with justice, security, and moral rules. Henry Sidgwick’s The Methods of Ethics first appeared in 1874 and offered a systematic examination of egoism, intuitionist morality, and utilitarianism. Sidgwick treated impartial benevolence as a central rational principle while acknowledging a persistent conflict between rational concern for one’s own good and rational concern for universal good. In Principia Ethica (1903), G. E. Moore separated the consequentialist structure of right action from hedonism, assigning intrinsic importance to goods such as personal affection and appreciation of beauty.\n\nThe term “consequentialism” acquired its modern prominence through G. E. M. Anscombe’s 1958 article “Modern Moral Philosophy.” Anscombe used it critically in discussing approaches that weakened the moral distinction between intended effects and foreseen side effects. Later philosophical usage broadened the term into a general category encompassing utilitarian and nonutilitarian theories. During the second half of the twentieth century, debates increasingly distinguished acts from rules, objective rightness from decision procedures, and welfare maximization from alternative accounts of value. Consequentialism thus became not merely another name for classical utilitarianism, but a framework within which competing theories could be formulated."
    },
    {
      "heading": "What makes an outcome good?",
      "text": "The most familiar consequentialist theories are welfarist: they maintain that the ultimate value of an outcome depends on the well-being of individuals. Welfarism does not by itself settle what well-being consists in. Hedonism identifies it with pleasure and the absence of pain. Preference-based theories instead emphasize the satisfaction of desires or preferences, sometimes restricted to preferences formed with adequate information and reflection. Objective-list theories recognize goods such as understanding, achievement, friendship, or autonomy even when these cannot be reduced to pleasurable experience or fulfilled desire.\n\nThese accounts can disagree sharply. Robert Nozick’s experience-machine example in Anarchy, State, and Utopia (1974) asks whether a life of apparently excellent experiences, produced artificially while a person remains disconnected from actual achievement and relationships, would be as good as a corresponding real life. The example challenges purely experiential accounts of welfare, not consequentialism as such. A consequentialist can answer that genuine relationships or achievement have value independently of experience. Similarly, preference theories must explain whether satisfying a misinformed desire benefits its holder and whether preferences concerning the suffering or subordination of others should enter moral evaluation.\n\nA separate issue concerns aggregation: how individual benefits and harms determine the value of an outcome as a whole. Total utilitarianism adds welfare across individuals, whereas average utilitarianism considers average welfare. Other consequentialist approaches give special importance to the distribution of benefits. Prioritarianism assigns greater moral weight to benefits received by people who are worse off, without necessarily treating equality itself as intrinsically valuable. Egalitarian consequentialism may instead count a reduction in inequality as one good among others. These positions belong to debates about distributive justice and should not be conflated: an outcome can be more equal without benefiting anyone, while a priority-weighted improvement can benefit the worse off without producing equality.\n\nConsequentialism can also be nonwelfarist and pluralist. It may include knowledge, artistic excellence, biodiversity, freedom, or just relationships among the goods to be promoted. Pluralism creates questions about comparison: how much environmental loss could be outweighed by gains in health, and are some goods incomparable? A theory may use weighted values, lexical priorities, or a partial ranking that leaves some alternatives neither better nor worse. Negative consequentialist theories give special or exclusive priority to reducing suffering or other bad outcomes. Their implications depend on whether preventing badness is balanced against creating goodness, and whether different forms of suffering can be aggregated. No single metric is built into consequentialism itself."
    },
    {
      "heading": "Acts, rules, motives, and decision procedures",
      "text": "Act consequentialism evaluates each available act by its consequences. Rule consequentialism instead evaluates rules or codes by the consequences of their acceptance, internalization, or observance, and then assesses actions through the selected rules. A rule against deception, for example, may be justified because general acceptance sustains trust, lowers the costs of communication, and protects people against manipulation. An individual lie could therefore be wrong under a consequence-selected code even if it would produce a slightly better result on that particular occasion.\n\nThe specification of rule consequentialism matters greatly. Evaluating universal compliance with a code is different from evaluating its realistic acceptance by people who sometimes misunderstand rules or fail to follow them. A code must also be learnable and usable. Excessively complex instructions may perform badly because teaching them is costly and agents cannot reliably identify their exceptions. Richard Brandt and Brad Hooker developed influential versions of rule-based consequentialist ethics that take such practical and psychological considerations seriously. On these accounts, rules are not merely convenient summaries of independently calculated act-level verdicts.\n\nRule consequentialism faces a structural dilemma. If a rule permits exceptions whenever breaking it produces the best consequences, the theory may collapse into act consequentialism. If it prohibits exceptions even when breaking it would be better, critics accuse it of “rule worship.” Defenders respond that the relevant justification concerns the consequences of accepting the whole code, including the effects of discretionary exceptions on judgment, trust, and compliance. A carefully specified code may contain emergency provisions without authorizing every agent to suspend rules whenever a local calculation appears favorable.\n\nA further distinction separates a criterion of rightness from a decision procedure. An act consequentialist can hold that right actions maximize good while denying that people should continually calculate outcomes. Habitual honesty, affection, professional standards, and reliable moral heuristics may produce better decisions than repeated explicit optimization. R. M. Hare’s Moral Thinking (1981) developed a two-level approach distinguishing intuitive moral thinking from critical reflection. Similar distinctions underlie indirect consequentialist defenses of motives and character. Such theories assess the consequences of cultivating dispositions that do not themselves direct attention constantly toward aggregate value. The resulting position is not automatically rule consequentialism: the ultimate criterion may still concern individual acts, even when the recommended psychology is rule-guided."
    },
    {
      "heading": "Actual consequences, expected consequences, and uncertainty",
      "text": "An action can produce a disastrous result despite being supported by the best available evidence. Conversely, reckless conduct can turn out well through luck. Consequentialist theories distinguish these cases by separating actual-consequence standards from evidence-relative or expectation-based standards. On an objective account, rightness depends on what an action actually brings about compared with its alternatives. On an evidence-relative account, the relevant comparison depends on the information available, or reasonably available, to the agent. These approaches need not assign the same role to rightness, justification, and blame.\n\nExpected-value consequentialism combines outcome values with probabilities. If \\(s\\) ranges over possible states, its basic calculation can be written:\n\n\\[EU(a)=\\sum_s P(s\\mid a)V(a,s),\\]\n\nwhere \\(P(s\\mid a)\\) represents the probability of state \\(s\\) under the choice of action \\(a\\), and \\(V(a,s)\\) is the value of that action’s outcome in that state. In causal decision settings, the probabilities must represent the effects of choosing an action rather than mere statistical associations. The theory then selects actions with the highest expected value. This model requires assumptions about probability, the comparison of values, and the treatment of risk. An ordinal ranking of outcomes alone is not enough to calculate expectations.\n\nMaximizing expected moral value should not be confused with maximizing an expected quantity such as money or years of life without further evaluation. The value assigned to outcomes may already reflect diminishing benefits, unequal burdens, or special concern for severe harm. Nor does uncertainty disappear once numerical probabilities are used. Long causal chains, limited evidence, dependence among events, and disagreement about the correct model can make precise estimates unreliable. Ambiguous probabilities, unbounded values, and extremely small chances of enormous benefits create further difficulties. Some consequentialists therefore explore robust decision rules, probability intervals, or restrictions on the kinds of speculative expectations that should govern action.\n\nThe distinction between right action and moral responsibility is particularly important here. An actual-consequence theorist can judge an action objectively wrong while holding that the agent was reasonable and not blameworthy. An expectation-based theorist can judge a lucky gamble wrong because its prospects were unacceptable when chosen. Both must address the costs of obtaining additional information: investigation itself consumes resources and can delay action. A consequentialist account of inquiry therefore considers not only what evidence would improve a decision, but whether acquiring it is worth its expected costs. These problems connect moral reasoning with decision theory and with the epistemology of responsible belief."
    },
    {
      "heading": "Rights, justice, and constraints on harming",
      "text": "A central objection to consequentialism is that maximizing aggregate value may permit sacrificing an individual for benefits to others. Familiar examples include punishing an innocent person to prevent unrest or killing one healthy person to distribute organs among several patients. Such thought experiments aim to isolate a conflict between producing better totals and respecting individuals as persons with claims against certain treatment. Their force does not depend on whether the imagined practices would normally succeed in real institutions.\n\nConsequentialists commonly respond that the full consequences of these practices include fear, error, abuse, damaged trust, and weakened institutions. A system permitting fabricated convictions would create incentives for misconduct and expose everyone to insecurity. Similarly, medical care depends on credible protections for patients. These replies can explain why consequentialist reasoning supports strong legal and professional safeguards. Critics nevertheless ask what the theory says when a case is deliberately constructed to remove such adverse effects. If violating a right really would produce the best outcome, a pure maximizing act theory appears committed to permitting or requiring the violation.\n\nThe trolley problem, associated with Philippa Foot’s work in 1967 and later developed by Judith Jarvis Thomson, examines distinctions among redirecting a threat, using someone as a means, and allowing harm to occur. Consequentialists may accept some of these distinctions because they reliably affect outcomes, or may include features of how outcomes arise within a richer evaluative account. Nonconsequentialist approaches often treat the distinctions as independently relevant. The doctrine of double effect, for example, distinguishes an intended harm from a foreseen but unintended side effect under specified conditions. Consequentialism does not ordinarily give that distinction decisive force merely because one effect was intended and another was not.\n\nSome consequentialist theories incorporate human rights or the fulfillment of rights directly into the assessment of outcomes. Others defend rights instrumentally or through consequence-selected rules. These approaches can yield extensive protections, but they differ over whether rights may be traded off against other goods and how violations should be counted. A related objection, developed prominently by John Rawls in A Theory of Justice (1971), concerns the separateness of persons: gains to some people do not compensate others in the way that a later benefit may compensate one person for an earlier cost. Consequentialists must therefore explain not merely why an aggregate is larger, but why that form of aggregation appropriately represents the claims of distinct individuals."
    },
    {
      "heading": "Demandingness, integrity, and personal relationships",
      "text": "Maximizing consequentialism can impose demanding obligations. If a person can produce more good by donating resources than by spending them on personal projects, the maximizing principle appears to require donation. The same reasoning can apply repeatedly, potentially leaving little room for leisure or optional generosity. Peter Singer’s 1972 essay “Famine, Affluence, and Morality” argued that preventing something bad is obligatory when it can be done without sacrificing anything of comparable moral importance. The essay sharpened a wider debate about the obligations generated by distant, preventable suffering and poverty.\n\nThe demandingness objection concerns the structure of duty, not simply reluctance to help. Ordinary morality distinguishes what is obligatory from supererogation: actions that are admirable but beyond duty. A strict maximizing theory appears to narrow that category because doing less than the best available act is wrong. Replies include accepting demanding obligations, distinguishing standards of rightness from appropriate blame, and developing less demanding forms of consequentialism. Satisficing consequentialism, defended by Michael Slote in the 1980s, permits actions whose consequences are sufficiently good even if they are not optimal. Such a theory must explain where the threshold lies and why foregoing further benefits is sometimes permissible.\n\nBernard Williams’s contribution to Utilitarianism: For and Against (1973) raised a related objection about integrity. His examples ask whether an agent must abandon deep commitments whenever doing so improves overall outcomes. The concern is not merely that calculation is unpleasant, but that the theory treats a person’s commitments as inputs to an impersonal production of value. Williams also criticized the tendency to make agents responsible for harms that others will commit unless they intervene in morally troubling ways. Consequentialists answer that commitments and personal agency can themselves be valuable, while critics question whether valuing them as outcome components captures their authority for the person concerned.\n\nFriendship and family relationships create a parallel problem of partiality. Parents ordinarily have special responsibilities to their children, and friendship involves concern for particular people rather than interchangeable recipients of benefit. Indirect consequentialists argue that stable special relationships promote care, development, and trust more effectively than constant impartial allocation. Peter Railton’s 1984 discussion of alienation defended the possibility of an objective consequentialist whose everyday motives are not organized around maximizing calculation. The remaining question concerns justification: whether a relationship is adequately respected when its ultimate defense rests on its contribution to overall good. Care ethics and other relational approaches often resist that explanatory priority even when they endorse many of the same practices."
    },
    {
      "heading": "Population, future generations, and nonhuman life",
      "text": "Consequentialist reasoning extends naturally to people separated from an agent by distance or time, provided that their welfare is included in the theory’s evaluative scope. This raises difficult questions in population ethics, where choices affect not only how well people live but also how many people exist and who they are. A policy that improves the lives of existing people differs from one that brings additional people into existence. Ordinary comparisons of benefits cannot simply assume that creating a happy person benefits that person in the same way that relieving an existing person’s suffering does.\n\nDerek Parfit’s Reasons and Persons (1984) made these difficulties central to contemporary ethics. His “Repugnant Conclusion” challenges total welfare theories: for any population enjoying very high quality of life, there could be a sufficiently larger population whose members have lives only barely worth living but whose total welfare is greater. Average welfare theories avoid that particular implication but face others, including the possibility of opposing the addition of a flourishing person merely because that person’s welfare falls below an already very high average. Critical-level and person-affecting approaches offer alternatives, each with further theoretical costs.\n\nParfit’s nonidentity problem concerns actions that change which people will later exist. A long-term policy may alter migration, relationships, and conception times, so that different individuals are born under different policies. Someone born under a harmful policy may still have a life worth living and might never have existed under the better alternative. It is therefore difficult to explain the policy’s wrongness solely by claiming that it made that particular person worse off. Impersonal consequentialism can condemn a worse future without identifying an individual who would have been better off, whereas person-affecting views must formulate a more complex explanation. These issues are relevant to intergenerational equity and climate change.\n\nThe moral scope of consequentialism also extends beyond human beings in many versions. Bentham’s discussion of animals emphasized their capacity to suffer rather than their capacity to reason or speak. Singer’s Animal Liberation (1975) developed an influential argument for equal consideration of comparable interests across species. Equal consideration does not imply identical treatment: different capacities and needs affect what benefits or harms a being. Questions remain about which organisms are sentient, how to compare different forms of experience, and whether ecosystems or species have value independently of individual welfare. Welfare-based consequentialism, biodiversity-centered views, and broader environmental value pluralism can consequently recommend different policies even when they share a commitment to evaluating outcomes."
    },
    {
      "heading": "Public institutions, practical evaluation, and moral boundaries",
      "text": "Consequentialist reasoning is prominent in public policy because institutions routinely compare the effects of alternative uses of limited resources. Decisions about transport, environmental protection, taxation, and public health involve opportunities foregone: spending resources on one program prevents their use elsewhere. Cost–benefit analysis attempts to express diverse effects in comparable monetary terms, while cost-effectiveness analysis compares costs per specified outcome. These are evaluative methods rather than complete moral theories. Their conclusions depend on which effects are included, how they are measured, and whether distributional or rights-based limits are imposed.\n\nHealth policy illustrates both the usefulness and the limits of aggregate evaluation. A quality-adjusted life year combines duration of life with a weighting for health-related quality of life. Comparing costs per such unit can inform the allocation of healthcare resources, but it does not by itself establish a morally correct allocation. Decisions may also consider severity of illness, equal access, the position of disadvantaged groups, and procedural fairness. Disagreement can arise over whether a measure adequately represents benefits to people with disabilities or chronic conditions. These are disagreements about value and justice, not merely disputes over technical accuracy.\n\nIn research ethics, anticipated social benefit is important but does not normally replace requirements concerning informed consent and participant protection. The Belmont Report of 1979 articulated respect for persons, beneficence, and justice as distinct principles. This illustrates the difference between institutions that consider consequences and institutions committed exclusively to consequentialism. Similarly, legal systems can justify punishment partly through deterrence and public protection while maintaining independent restrictions on conviction, proportionality, and treatment. Consequentialist analyses must assess the behavior of real institutions, including mistakes and incentives, rather than assuming that officials can reliably identify and implement the best outcome.\n\nApplications to artificial intelligence expose another limitation of simple optimization. A system that maximizes a numerical target is not thereby implementing a defensible consequentialist ethics: the target may omit relevant harms, misrepresent human interests, or reward manipulation of the measurement process. Conversely, adding more outcome measures does not settle questions about consent, legitimate authority, or who may decide among conflicting values. Consequentialist contributions are strongest when they distinguish empirical predictions from normative commitments and make opportunity costs visible. Their practical use still requires specification of the affected parties, defensible comparisons among benefits and harms, and institutions capable of acting under uncertainty without treating a convenient metric as a complete account of moral value."
    }
  ],
  "links": [
    "ethics",
    "utilitarianism",
    "deontology",
    "virtue-ethics",
    "causation",
    "counterfactual",
    "david-hume",
    "george-edward-moore",
    "distributive-justice",
    "knowledge",
    "expected-value",
    "probability",
    "epistemology",
    "thought-experiment",
    "trolley-problem",
    "doctrine-of-double-effect",
    "human-rights",
    "john-rawls",
    "poverty",
    "supererogation",
    "intergenerational-equity",
    "climate-change",
    "public-health",
    "cost-benefit-analysis",
    "quality-adjusted-life-year",
    "informed-consent",
    "belmont-report",
    "artificial-intelligence"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}