In philosophy and argumentation, a reason is a consideration that counts in favor of a conclusion, action, attitude or fact, or that explains why something is so. Reasons typically answer a why? question and are often introduced by expressions such as because, since, as, in virtue of, or in order to. They are central to accounts of practical reason, epistemic justification, moral evaluation, and everyday explanation, and they figure prominently in law and deliberative discourse.
Philosophers commonly distinguish three roles for reasons. Normative (or justifying) reasons are considerations that count in favor of responding one way rather than another (e.g., that it is raining is a reason to take an umbrella). Motivating reasons are the considerations in light of which an agent actsâÂÂwhat the agent treats as counting in favor at the time, whether or not it in fact does. Explanatory reasons cite what explains an event or action; when agents are involved, these often refer to psychological states (for example, that someone believed they were late explains why they ran).
Debates concern what reasons are and how they work. Some hold that normative reasons are facts (or true propositions) rather than mere beliefs; others link them more closely to an agent's perspective. Reasons are said to play both a deontic role (helping to determine what one ought, may, or must do) and a deliberative role (serving as appropriate inputs to sound deliberation and, when taken up, becoming motivating reasons). Further questions include whether acting rightly must be done for the right reasons to have moral worth, and how normative and motivating reasons are related when guidance is difficult (for example, in surprise-party or akratic cases).
The literature also distinguishes epistemic reasons (which count in favor of believing a proposition) from practical reasons (which count in favor of actions or attitudes), and asks whether there is a unified treatment of bothâÂÂe.g., by understanding reasons as a kind of evidence. Disputes about the source of practical reasons are framed as internalism versus externalism: internalists tie a person's reasons to their actual or idealized motivational set (sometimes in broadly Humean terms), while externalists allow that there can be reasons independent of an agent's present motivations. A separate contrast between agent-neutral and agent-relative reasons concerns whether the content of a reason makes essential reference to the agent (as with special obligations to one's child) or not (as with impartial welfare considerations).
Because agents often face multiple, context-sensitive considerations, contemporary work analyzes how reasons are weighed, defeated, or enabled. Proposals address holism about reasons, pairwise and contrastive frameworks for permissibility, possible incommensurability or parity among options, and distinctions between justifying and requiring strengths. Related topics include exclusionary (higher-order) reasons that bar acting for certain first-order reasons, and questions about aggregating overlapping considerations. These issues connect the concept of a reason to broader discussions in ethics, rational choice, epistemology, and the theory of argument.
Many contemporary accounts start from a threefold classification.
Normative reasons play at least two roles. In their deontic role, they help settle what one ought/must/may do by weighing reasons for and against actions. In their deliberative role, they are the appropriate inputs to sound deliberation and can become motivating reasons when an agent responds to them.
A common view holds that normative reasons are facts, not merely beliefs; some identify them with true propositions, others with worldly states of affairs. Authors also distinguish between objective reasons (the facts that count for/against an act) and subjective or apparent reasons (considerations within an agent's epistemic perspective that make acting seem reasonable).
Competing characterizations include:
Reason-weighing is often modeled by metaphors of weight/strength, but holism, context-sensitivity, and incommensurability complicate simple additive pictures. Some argue for epistemic constraints on the reasons relevant to what one ought to do (the perspectivist idea that only what falls within one's perspective can determine one's ought), while others retain an objective ought sensitive to all the facts. Classic puzzle cases (e.g., unforeseeable harms) illustrate the tension.
Philosophers distinguish an agent's motivating reason from the fact that explains their action. For example, Othello kills Desdemona for the (false) consideration that she was unfaithful (motivating reason), while what explains his action includes that he believed she was unfaithful (explanatory reason).
Debate centers on whether motivating reasons are:
A widely discussed constraint says that one must be able to act for a good reason; this pressures psychologism if mental states as such cannot be good reasons. There is also debate over the epistemic condition for acting in the light of a factâÂÂsome argue that mere true belief is insufficient and that knowledge is required; others resist this requirement.
On guise of the good/guise of reasons views, acting for a reason involves taking a consideration as favoring the action (seeing something to be said for it). Alleged counterexamples (anger, akrasia, evil be my good) have prompted refinements (e.g., appealing to how things seem good at the time).
A natural idea is that normative reasons are supposed to guide us. This is developed as: a deliberative condition (we ought to be moved by them), an ability condition (they must be able to move us), or a combined condition (we can be moved by them via sound deliberation). Puzzles about self-effacing reasons (e.g., surprise-party cases) challenge simple guidance theses and motivate careful distinctions between deontic and deliberative roles.
Sometimes one does the right thing but not for the right reason (e.g., rescuing a child for praise). Many hold that moral worth/credit requires acting for the very reasons that make the action rightâÂÂor from the right kind of concern or normative knowledge.
Philosophers distinguish capacities of theoretical (epistemic) and practical reason. Epistemic reasons (evidential reasons) count in favor of believing a proposition; practical reasons count in favor of actions or attitudes. Some propose a unified treatmentâÂÂe.g., that reasons are a kind of evidence across domains. In argumentation, a reason may be a premise (or set of co-premises) supporting a conclusion; explanatory reasons can clarify how something could be true without by themselves showing that it is true.
In debates about practical reason, philosophers ask whether there is a necessary connection between an agent's normative reasons and the agent's motivation. Reasons internalism (a family of views) holds that for a consideration to be a reason for an agent, it must stand in some privileged relation to facts about that agent's motivation or motivational psychology; reasons externalism denies any such necessary connection. Although framed about action, these positions bear on broader issues about normativity, rationality, and moral obligation.
Scholars distinguish several axes along which internalist theses vary.
A prominent actual-state thesis is the Humean Theory of Reasons (HTR): one has a reason to perform an action only if doing it would serve some actual desire of one's own. It is commonly paired with the Humean Theory of Motivation (HTM) that desires are necessary and beliefs not sufficient for motivation. Critics argue that HTR conflicts with two widely held assumptions: moral rationalism (that moral wrongness entails a reason not to act) and moral absolutism (that some acts are wrong for anyone). Together these suggest there are reasons independent of an agent's actual desires. Responses include rejecting absolutism or rationalism (e.g., forms of relativism or error theory), or weakening internalism to counterfactual forms that do not depend on an agent's present psychology.
Critics argue internalist theses can undergenerate reasons: e.g., we seem to have prudential reasons now (to study for a trip we'll later regret being unprepared for) even if we presently lack supporting desires; internalists reply with overdetermination strategies or by challenging the data. Conversely, simple desire-based views appear to overgenerate (mud-drinking, blade-counting); internalists often treat desire as a necessary but not sufficient condition, or else appeal to background standards filtering which desires yield reasons.
Recent work explores whether counterfactual internalisms can avoid triviality without reverting to actual psychology, and whether externalist accounts can explain both categorical moral reasons and the everyday desire-sensitivity of many reasons; anthologies collect competing approaches and clarifications.
Scholars often distinguish between agent-neutral and agent-relative normative reasonsâÂÂtwo patterns by which considerations favor actions. On a classic principle-based view, a reason is agent-neutral when the associated principle contains no non-trivial reference to the agent for whom it is a reason; it is agent-relative when such reference is essential to the reason-giving condition. Thus, impartial welfare-maximizing considerations are standardly treated as agent-neutral, whereas egoistic reasons or special-relationship reasons (e.g., to help one's own child) are agent-relative.
Other ways of drawing the line have been proposed. On a reason-statement approach, a reason is agent-relative iff a full statement of that reason must use pronominal back-reference to the very agent for whom it is a reason (e.g., "that it is in his interest"). A different, perspective-based approach ties neutrality to whether reasons can be recognized from suitably objective standpoints, and takes some reasons (e.g., nearest-and-dearest) to be appreciable only from more first-personal perspectives. Each formulation faces pressures: the reason-statement view depends on contentious assumptions about irreducibly indexical facts, while the perspective view makes classification hinge on controversial theses about objectivity.
Many hold that reasons are holistic and context-sensitive: what counts in favor in one case may be silenced or reversed in another by defeaters; some background conditions are mere enablers rather than reasons themselves. To accommodate this holism while retaining the clarity of the principle-based distinction, some theorists use default principlesâÂÂhedged generalizations that say, roughly, "F is a reason for p to àunless some feature of the situation explains why not." On this picture, neutrality vs relativity turns on whether the reason-clause contains a non-trivial free agent variable (e.g., "ps pleasure"). Objections that such principles are vacuous are answered by noting that the "feature of the situation" is constrained to be contingent, not any necessary truth, so the hedging does real explanatory work rather than trivializing the claim.
Several nearby distinctions are sometimes conflated with the agent-neutral/agent-relative divide:
The distinction structures debates about consequentialism and deontology. Deontic constraints can be modeled via agent-relative (often time-relative) teleological reasons, a strategy used to "consequentialize" deontological theories in structural terms. It also highlights surprising asymmetriesâÂÂe.g., maximizing expected utility can introduce agent-relativity via whose expectations count. The distinction has been central in work on expressivism about normativity, and in Kantian attempts to vindicate deontic prohibitions without agent-relative value.
Many theorists analyze how multiple considerations interact to determine what is permissible, required, or best by appealing to the weighing of reasons. On this approach, reasons function as direct contributors to a verdict, while various contextual features can indirectly enable, disable, amplify, or attenuate the relevance of those reasons.
A widely discussed templateâÂÂsometimes called Monist Pairwise PermissibilityâÂÂsays an option àis permissible iff, for each alternative, the reasons for àare not outweighed by the reasons for that alternative. This treats permissibility as a "tournament" of pairwise comparisons rather than a free-for-all among all reasons at once. Within this framework, some defend contrastive reasons, whose weight (or applicability) varies with the specific alternative under consideration (e.g., a reason for àagainst option A may not be a reason for àagainst option B). Contrastivism has been developed for practical and epistemic cases and used to explain puzzling triads like Kamm's "Café or Kid" case, where what competes with promise-keeping changes once a third option (saving a child at severe personal cost) is introduced.
On one influential view, talk of a "reason against" ÃÂ just is talk of a "reason for" some specified alternative to ÃÂ in the same pairwise contest; which locution we use depends on explanatory emphasis. How we individuate options matters to weighing. Maximalism holds that reasons fundamentally apply to maximal, mutually exclusive and exhaustive options (fully specified acts), and that permissibility for coarser options derives from verdicts about these maximal ones.
Against atomism (which treats a consideration's status/strength as invariant across contexts), holism claims that context can change whether something is a reason at all (via enablers/disablers) or how weighty it is (via amplifiers/attenuators). Examples include coerced promises (disablers) and special relationships (amplifiers of reasons to aid one's child).
To explain widespread and stable cases where more than one option is permissible (moral "options"), some reject strict trichotomy (heavier/lighter/equal) and allow parity: two options can be comparable without being precisely equal; small improvements need not break parity. Others worry parity alone cannot preserve permissible partiality or supererogation if permissibility is still pairwise and monistic.
An alternative is Weight Pluralism: reasons have at least two distinct weightsâÂÂjustifying (good at making an action permissible) and requiring (good at making alternatives impermissible, hence making the action required). On the Pluralist Pairwise view, àis permissible iff ÃÂ's justifying weight is not outweighed by rivals' requiring weight. This promises stable, ubiquitous options and a tidy explanation of permissible partiality and supererogation.
Some pluralists rank betterness among permissible options by requiring weight ("Simple Betterness"), while others posit a third, "favoring/merit-conferring/erogatory" dimension to handle "ought-as-best."
On one usage, ought tracks the best permissible option, not merely the required one; this separates ranking from permissibility and helps locate supererogation (doing better than some permissible alternative without being required). Competing proposals ground this ranking in requiring weight or in a distinct favoring/merit dimension.
Beyond first-order weighing, some defend exclusionary reasonsâÂÂhigher-order reasons not to act for certain first-order reasons (e.g., a promise to ignore self-interest in a decision). Critics worry this divorces "all-things-considered" action from the first-order balance, but defenders argue such reasons structure rational deliberation in law and practical life. In epistemology, some interpret certain higher-order evidence (e.g., risk of reasoning impairment) as excluding one's first-order evidence from justifying belief even if it remains evidence in a purely evidential sense.
Simple additive models face two challenges: (i) derivative reasons (e.g., heat and rain each count against running, but together they may be less bad) and (ii) overlapping fundamental reasons (e.g., "pain" and "severe pain" shouldn't double-count). Contemporary views often restrict addition to non-overlapping fundamental reasons for maximal options, holding fixed the relevant conditions/modifiers.
Sometimes morality, prudence, and epistemic rationality deliver conflicting verdicts. One family of views posits a supreme perspective whose weighing settles what one ought to do simpliciter; others endorse normative pluralism: several perspectives have authority but no single, integrated court of appeal. The literature explores whether differences between (say) altruistic and self-interested reasons are best handled by parity, pluralist weights, or pluralism about perspectives.