Intentional Action

The Philosophical Problem of Intentional Action

Intentional action is a form of doing in which the agent's conduct is intelligibly directed toward something the agent takes itself to be doing. The governing philosophical problem is to identify the relation that makes a performance intentional: an explanatory relation between action and reason, a causal relation between action and intention, an epistemic relation between an agent and its performance, a constitutive relation between action and practical understanding, or a structure of guidance sustained while the action unfolds. These candidates are connected. Their differences become decisive in examples where a person has a reason but acts for another reason, achieves a desired result by accident, executes a plan without attending to each movement, or knowingly performs one action while inadvertently producing another outcome.

Consider an archivist who turns a lever to open a window for ventilation. The movement also activates an alarm unknown to the archivist. The occurrence can be accurately described as moving the lever, opening the window, and activating the alarm. The first two descriptions express what the archivist is deliberately doing, while the third describes an unanticipated consequence. If an external mechanism had instead moved the archivist's arm, identical physical displacement would occur without the same exercise of agency. If anxiety caused an involuntary tremor that moved the lever and opened the window, the desired result would occur without being accomplished through intentional performance. An adequate theory must distinguish these patterns without defining action merely by success.

The philosophical tradition does not possess one universally accepted set of necessary and sufficient conditions for intentional action. Nonetheless, a recognizable research program emerged through Aristotle's account of voluntariness and choice, medieval discussions of willing and intention, early modern arguments about motives and the will, and twentieth-century analyses of reasons, action descriptions, causation, and practical knowledge. Contemporary debate concerns the coordination of those dimensions, especially over time and in embodied interaction (Aristotle, Nicomachean Ethics, III.1–5; Anscombe, 1957/2000; Davidson, 1963; Bratman, 1987; Schlosser, 2019).

This monograph advances a description-sensitive and guidance-centered synthesis. Intentional action requires a practical orientation toward a describable undertaking and an appropriate realization of that orientation through the agent's activity. The orientation may be explicit or dispositional; guidance may be conscious, skilled, or distributed across stages; and success is contingent. These claims will be defended through comparison with causal, teleological, practical-knowledge, and planning approaches, while marking the points at which further philosophical and empirical investigation remains open.

Definition, Scope, and the Logic of Action Descriptions

The relevant unit of inquiry is an agent's performance under a description. We can say that an occurrence E is intentional under description D for agent A when D identifies something A undertakes as part of A's practical project and E occurs through an appropriately guided realization of that undertaking. This formulation is an analytical proposal, not a report of a universally agreed definition. It keeps distinct three questions: whether something is an action at all, whether it is intentional under a particular description, and whether the intended outcome was successfully achieved.

The phrase 'under a description' matters because descriptions are ways of specifying what occurs. One bodily movement may be described by its immediate manner, its instrumental function, its social meaning, and its further effects. The truth of each description does not make it an equally apt characterization of what the agent intended. If I intentionally sign a receipt that transfers ownership of a manuscript and unknowingly makes a third party eligible for a prize, the latter effect may be a consequence of what I did but was never among the aims or understood means of the performance. This difference cannot be settled by measuring muscle movement.

The description-relative question should not be conflated with the metaphysics of how many events exist. A coarse-grained event theorist can regard signing, transferring ownership, and establishing eligibility as descriptions of one event. A fine-grained theorist can count distinct acts or property-instantiations. Both need a story about which aspect was practically represented or governed. Anscombe (1957/2000, §§6, 23) and Davidson (1971/1980) established the centrality of this issue, while Goldman (1970) supplied a different framework for act structure.

Intention here designates a practical orientation toward doing, rather than the broad philosophical sense of intentionality as directedness or aboutness. A belief about a storm has intentional content in philosophy of mind; it does not thereby constitute an intention to act. Searle (1983, chapter 3) investigates the relation between the content and conditions of satisfaction of intentions and actions. The present research concerns how that content is undertaken and realized in conduct, including cases in which an agent's practical orientation changes during execution.

Voluntary, Deliberate, Purposive, and Intentional

The terms voluntary, deliberate, purposive, and intentional intersect without naming a single property. Voluntariness concerns the originating conditions of action and the agent's relevant knowledge; deliberateness indicates prior consideration or choice; purposiveness indicates directedness toward an end; intentionality in the practical sense indicates that a performance falls under a description embraced within the agent's practical orientation. A voluntary action may be impulsive, an intentional action may be habitual, and an action may be purposive without an articulated deliberative episode.

Aristotle distinguishes voluntary action, hekousion, from choice, prohairesis. In Nicomachean Ethics III.1–3 he locates voluntariness in the agent as source and in knowledge of particulars, while describing choice as involving deliberation. It would erase Aristotle's argument to translate both terms as simply 'intentional action.' The distinctions permit anger-driven or desire-driven behavior to be voluntary while denying that all such behavior exemplifies reflective choice. The modern problem therefore cannot be projected backward as though Aristotle employed one contemporary English category.

A second distinction concerns intended consequences and foreseeable side effects. A surgeon can intentionally perform an incision while foreseeing pain; a pilot can intentionally reroute a flight while foreseeing a delay. Foreseeability changes the deliberative and moral context but does not, without further argument, make every anticipated outcome an intended end. Bratman (1987, chapters 9–10) develops the practical difference between intending an effect and expecting an effect, especially where expectations do not settle an agent's commitments in the manner that plans do.

Finally, a person's intention can be present while particular movements escape attention. The pianist does not choose each contraction separately; a fluent speaker does not formulate a fresh explicit plan for every syllable. Skilled performance exhibits structured activity under an endorsed undertaking. Attention is one possible form of guidance rather than the measure of intentionality. Conversely, repeatedly rehearsing an imagined action or wishing it occurred can supply mental content without instituting a performance. The philosophical task is to explain the transition from practical orientation to unfolding execution.

Aristotle: Voluntariness, Choice, and Practical Reason

Aristotle supplies a foundational vocabulary for investigating action even though his principal question concerns ethical character and responsibility. In Nicomachean Ethics III.1 (1109b30–1111b3), he distinguishes acts done under compulsion or through ignorance from voluntary acts and tests mixed cases in which an agent acts under severe pressure. The source of motion must be understood within the concrete practical situation. In III.2 (1111b4–1112a17), he argues that choice is narrower than voluntariness: animals and children can act voluntarily, while choice implicates deliberation of a distinctive kind. In III.3 (1112a18–1113a14), deliberation concerns what lies within the agent's capacity and how ends can be realized.

The structure of Aristotle's analysis prevents the modern philosopher from using deliberation as the universal mark of agency. Someone who suddenly catches a falling glass can respond knowingly and purposefully before undertaking explicit calculation. Yet Aristotle also explains why complex activities are answerable to a rational ordering of means. Practical reason moves among things that can be done, circumstances that matter, and feasible steps. The agent's action is intelligible through that organization, not merely through a subsequent causal narrative.

The distinction between praxis and poiēsis, action and production, also matters. Some activities produce a separable outcome, while others realize their point in the doing. An action of keeping a promise may involve writing a letter, but fidelity to the promise is a normative description of the practice rather than an isolated physical product. Aristotle's account in Nicomachean Ethics VI associates phronēsis with deliberation about living well and makes the practical character of judgment visible.

The classic difficulty is that voluntariness, choice, ethical virtue, and contemporary intentionality answer overlapping but different questions. The advantage of Aristotle's framework is a stratified account of action: something originates in an agent, is undertaken with relevant awareness, may involve choice, and can express a stable practical character. The limitation, for our present inquiry, is that it does not by itself articulate the modern causal deviance problem: a state of desire could produce a suitable outcome through a causal route that bypasses the agent's performance.

Medieval Intentio and the Ordering of Means to Ends

Medieval philosophical reflection developed the relation between will, knowledge, and the direction of conduct. Thomas Aquinas's Summa Theologiae, I–II, question 6, analyzes the voluntary and involuntary, including violence, fear, desire, and ignorance. Question 12 asks whether intention belongs to intellect or will and how intention is ordered to an end. Question 13 analyzes choice. The sequence reveals a structured account in which cognition presents objects and possibilities, willing is ordered toward goods, and deliberative selection concerns means. Aquinas's distinctive theological commitments form part of his original setting; the logical distinctions remain important for action theory.

An intention to achieve an end cannot be identified simply with representing the end as desirable. The agent must orient some possible performance toward that end. Aquinas's treatment makes available a distinction between an act's ultimate end, an intermediate end, and the means chosen here and now. An intention can organize a chain of activities without requiring that the agent currently inspect every subordinate movement. An instrument can be chosen for an intended purpose even when its ordinary operation proceeds through learned competence.

This tradition also motivates a question about practical knowledge. The agent's understanding of what is being done plays a productive role in organizing the deed. Such knowledge is different in direction from a detached spectator's discovery of an already completed fact. Anscombe's later revival of practical knowledge explicitly places herself in conversation with this Aristotelian and Thomistic inheritance (Anscombe, 1957/2000, §§32, 48; Aquinas, Summa Theologiae, I–II, qq. 6, 12–13).

Historical reconstruction requires restraint in translation. The Latin intentio has meanings internal to scholastic theories of intellect and will, and modern 'intention' has developed additional psychological and linguistic functions. Their genealogical relation is philosophically illuminating precisely because the theoretical systems differ. A substantial historical line runs from ordered willing and practical understanding to modern accounts of plans and rationalizing explanations, while leaving room for disagreement over whether intending is a separate mental state, an activity of willing, or a constitutive aspect of action itself.

Early Modern Motives, Will, and Rational Agency

The early modern problem of action reorganized classical concerns around the will, passions, reasons, and the explanation of conduct. David Hume argues in A Treatise of Human Nature, Book II, Part III, section 3, that reason alone cannot supply the motivating force that produces an act. His point depends on distinguishing cognition of relations or facts from the passions that move us. He challenges any model of rational agency that treats an intellectual judgment, by itself, as sufficient to explain motivation. A later causalist can inherit this explanatory division while rejecting Hume's more comprehensive psychology.

The challenge remains acute in contemporary theories: a person can judge that exercise is desirable and remain unmoved; another can recognize a reason for honesty yet act from fear; a third can act purposefully without rehearsing an evaluation. We must distinguish judging there to be a reason, being moved by a consideration, intending an act, and successfully carrying it out. Explanations which compress all four into one 'decision' lose the phenomenon they are meant to explain.

The modern concept of the will also carries distinct normative commitments. The question of whether conduct is self-determined or imputable is connected to the analysis of moral duty, autonomy, and responsibility. Kant's moral philosophy asks whether the maxim of action can be adopted as a principle and how a rational agent determines its will through reasons. That concern connects action with imputability, though the present monograph does not equate Kantian practical lawgiving with the ordinary-language predicate 'intentional.'

A productive conclusion of this history is methodological. The account of an action's grounds, its motivational efficacy, and its normative assessment are independent tasks that interact. A bad motive can explain an intentional act; an exemplary reason can fail to motivate; an intention can be executed incompetently. The history from Aristotle through Hume motivates the modern analysis of action as a layered phenomenon rather than the output of a single mental item.

Ryle, Wittgenstein, and the Grammar of Doing

Twentieth-century analytic philosophy inherited disputes about the will and tested them against ordinary descriptions of action. Gilbert Ryle's The Concept of Mind (1949) criticized the image of a hidden inner performance standing behind every intelligent outward performance. If an intelligent deed required a prior intelligent act of willing, the explanation would invite a regress. Practices, dispositions, skills, and criteria for describing conduct become central to understanding what an agent does.

Wittgenstein's Philosophical Investigations §621 asks what remains when the fact that one's arm rises is subtracted from one's raising it. The question exposes the difficulty of identifying action by appending an invisible occurrence to a bodily event. A description of the raised arm identifies a movement; the description 'I raise my arm' implicates the subject's doing. The task is to understand that difference without assuming the answer resides in an inaccessible introspected episode.

These arguments encourage attention to the publicly intelligible pattern of a performance. A request for reasons operates in practices of explanation and accountability. The agent's competence in proceeding, correcting, and making sense of the undertaking informs whether we count an occurrence as something done. This orientation need not exclude causal analysis, but it resists treating a hidden volition as the automatic criterion of action.

There is an important limit to purely grammatical analysis. Two events may support the same ordinary description while differing in their causal history: opening a door intentionally, opening it by an uncontrolled reflex, or opening it through a motor disturbance. Language gives us distinctions worth explaining; it does not always settle the underlying mechanism. The postwar literature on reasons and causes developed as an attempt to preserve the intelligibility of intentional explanation while clarifying what makes an explanation true of this performance rather than another.

Anscombe's Intention: The Special Question Why

G. E. M. Anscombe's Intention (1957; second edition 1963) established a modern agenda by examining expressions of future intention, intentional action, and the intention with which one acts. In §5 she characterizes intentional actions through a special application of the question 'Why?': one asks for the agent's reason for doing the act, and an answer can disclose its practical intelligibility. This is not an invitation to posit a mental sentence that accompanies every movement. Anscombe's investigation is grammatical, epistemological, and practical at once.

The question distinguishes kinds of answers. A person may explain a startle by citing a sudden noise while lacking a reason for startling. The noise can be a cause of what happened, yet the explanation does not characterize the movement as an action undertaken for an end. Conversely, an agent can explain walking toward a station by the desire to catch a train; the latter answer places the walking within a practical ordering. Anscombe carefully separates causes, motives, and intentions in her initial sections.

In §§6–8 she develops the importance of knowing what one is doing under a description. If someone knowingly opens a valve but does not know it activates a sprinkler, the agent can be intentional in opening and unintentional in activating. The predicate 'intentional' is thus description-sensitive. The special question does not supply a simple mechanical test: people may lie, forget, be mistaken, or give retrospectively embellished explanations. Its value is that it identifies a distinctive mode of intelligibility.

A charitable interpretation of Anscombe recognizes that intentional action is a structured practical phenomenon in which undertaking, description, and knowledge are intimately related. The open issue concerns whether the agent's knowledge is a condition of intentionality in every case or whether intentional action can include mistakes, incomplete awareness, and failures of self-interpretation. That issue became central to later disputes between practical-knowledge approaches and event-causal theories (Anscombe, 1957/2000, §§5–8, 23, 32, 48; Setiya, 2008).

The Pumping Example and Multiple Levels of Action

Anscombe's extended pumping example in Intention §23 shows how one performance can be placed within successive descriptions. A person moves an arm, operates a pump, replenishes a water supply, and, in the case she describes, participates in poisoning it. Questions about why each stage is undertaken elicit further ends and reasons. The philosophical point concerns the order of practical intelligibility, the agent's knowledge, and the way different descriptions attach to the activity.

We can adapt the example without preserving its morally disturbing details. Suppose a caretaker rotates a crank to fill a reservoir in order to sustain an orchard. The caretaker knows that turning the crank powers a pump and that the pump fills the reservoir; its operation also awakens a nearby security system, which the caretaker does not know. The same continuous performance therefore lies under descriptions with different intentional statuses. The goal of caring for the orchard structures the intermediate activity; the unexpected alarm does not become intended merely because it is caused.

This structure yields an important distinction between a means, an end, and an effect. The caretaker intends the intermediate operation as a means to reservoir filling. The caretaker intends reservoir filling as a means to irrigation. An external side effect is caused by the action without being included in the practical project. Even known side effects demand further analysis: an agent may recognize that something will happen and proceed while neither adopting it as a purpose nor relying on it as an instrumental step.

Action descriptions also matter to social attribution. 'Moving a hand,' 'signing a contract,' and 'transferring authority' can describe aspects of a single performance. The social and institutional predicates depend on conventions, documents, and relations; they need not correspond to additional bodily motions. The ontology of action must therefore accommodate physically grounded activity and relational descriptions while preserving what the agent is answerable for doing intentionally.

Davidson: Reasons as Causes

Donald Davidson's 'Actions, Reasons, and Causes' (1963) argues that the reasons by which an agent's conduct is rationalized also explain it causally. His 'primary reason' comprises a favorable attitude toward actions of a certain kind and a belief that the particular undertaking has the relevant character. Where the explanation cites a primary reason, it represents the act as attractive or worthwhile from the agent's perspective. Davidson's decisive step is to insist that such a reason be related to the act as a cause if it is to explain why the action occurred.

The argument begins with explanatory competition. Suppose a writer has several reasons to visit a library: to consult a manuscript, meet a colleague, or enjoy quiet. The writer visits because of the manuscript. Simply enumerating all attractive features cannot identify why this visit happened. If explanatory relevance is historical, the actual reason must connect to the occurrence in the way that makes it happen. Causal explanation promises that connection. Davidson's thesis is stronger than the assertion that an action has some cause; it requires the cited rationalizing reason to be causally efficacious.

Davidson's reconstruction protects intentional explanation from a separation between rationality and causality. Reasons belong to the agent's understanding and can be causes of its activity. Nevertheless, Davidson does not demonstrate that every true rationalization can be reduced to descriptions of brain events or that all action must be explained through one pair of mental states. His work is frequently read as founding the standard causal theory while remaining more nuanced about reductive ambitions and causal deviance (Davidson, 1963; Piñeros Glasscock & Tenenbaum, 2023).

The principal difficulty concerns the 'right way' in which reasons must cause action. A reason may produce a bodily result by generating nervousness, distraction, or an uncontrolled response. This establishes a causal relation, yet the resulting event may fail to execute the agent's practical undertaking. Davidson's theory identifies an indispensable distinction between actual and merely available reasons, but causal history alone must be supplemented by constraints on how intentions produce or guide performance.

Intentions Beyond Belief and Desire

Belief and desire are not interchangeable with intention. A belief represents what the agent takes to be the case; a desire presents an attractive outcome or condition; an intention settles, at least provisionally, what the agent is to do. An agent can desire two incompatible outcomes, believe several routes feasible, and yet intend to pursue one. The difference becomes visible in sustained projects where a practical commitment constrains subsequent deliberation.

The proposition 'I will work on this manuscript tomorrow' can express prediction, hope, or intention depending on how it functions in practical thought. If it is a settled intention, the agent can order resources, exclude competing plans, and treat certain further questions as answered unless circumstances warrant reconsideration. Intentions also impose demands of means-end coherence: one cannot intelligibly sustain the commitment to travel while rejecting every known necessary means as a matter of policy.

This distinction should not imply that an intention is always conscious, verbally formulated, or carried by a separate mental episode. Someone can maintain an intention through dispositions to prepare, adjust, and proceed. The representational content matters because it determines what counts as continuation, success, or abandonment. Yet the form of storage and rehearsal may vary with expertise and temporal distance.

Searle (1983, chapter 3) distinguishes relations among intentions, actions, and conditions of satisfaction; Bratman (1987) gives the most influential account of intentions as elements of plans; and Mele (2003) examines the motivational basis of agency. Their arguments show why intention cannot be identified simply with an especially strong desire. Desire can persist when conduct is decisively settled elsewhere; intention is distinguished partly by its place in the organization of performance.

Bratman's Planning Theory and Temporally Extended Agency

Michael Bratman's Intention, Plans, and Practical Reason (1987) treats intentions as elements of partial plans that coordinate activity over time. Human agents have limited cognitive resources and cannot reconsider every practical question before each action. Plans organize future behavior by selecting goals and means, supplying a framework for further deliberation, and supporting coordination with other people. Intentions consequently have a stability that is defeasible rather than absolute.

Imagine an investigator preparing an archive over several months. The investigator may intend to organize documents, schedule digitization, commission translations, and publish a catalog. Different acts unfold on different days. Their intelligibility depends on a continuing practical commitment connecting temporally separated performances. No single initial neural event adequately describes the internal structure of this extended project; equally, an indefinitely repeated deliberation would make ordinary action impracticable.

The planning approach clarifies the distinction between intending an outcome and merely expecting it. An agent can expect delay as a side effect of a necessary procedure while maintaining an intention to complete the task. Expected effects enter risk assessment and moral responsibility, but they need not guide means selection as intended elements. Bratman's chapters 8–10 explore intention's dual role in planning and acting, together with the intention/expectation distinction.

Planning theory faces its own boundary questions. Agents frequently execute spontaneous acts without explicit future commitments; skilled agents also depart flexibly from plans. Bratman's framework is strongest when interpreted as identifying a distinctive dimension of intelligent agency, rather than demanding that every intentional movement result from a prior articulated plan. A complete theory needs to explain how long-term commitments are implemented through situated, revisable, and sometimes improvisational guidance.

Intention in Action and the Execution of Plans

An intention about the future and an intention in action differ in temporal role. The first organizes prospective conduct, while the second is exercised in the ongoing performance. Searle's Intentionality (1983, chapter 3) emphasizes the relation between intention and actual doing; the content of intention has conditions of satisfaction tied to the agent's enactment. The idea makes it possible to describe intentional action as something more than an external sequence following a previously settled mental state.

The distinction matters in failure cases. A person can genuinely intend to draw a circle but produce an irregular figure. The intention remains practically operative because it directs the drawing, prompts correction, and determines the kind of failure. An observer who sees only the final shape may miss the undertaking that was being attempted. Conversely, a person can make a visually perfect circle by accidental machinery without having enacted an intention to draw it.

A theory of intention in action should allow feedback and reorganization. During a walk toward a meeting, a blocked road can lead the agent to take another route without abandoning the overarching project. The agent continues acting intentionally under the description 'going to the meeting' while revising the means. Guidance persists across transitions between anticipation and activity, without requiring infallible awareness of every step.

The most difficult question is how a present intention relates causally or constitutively to ongoing movement. A mental-state theory may identify a control architecture through which planning states shape behavior; a practical-knowledge theory may describe the agent's ongoing understanding as internal to doing. The synthesis defended here treats both as constraints on a unified phenomenon: the undertaking specifies what is being done, and the activity is answerably organized toward that specification.

Action Guidance, Control, and the Frankfurt Objection

Harry Frankfurt's 'The Problem of Action' (1978) criticizes accounts that distinguish an action from a bodily happening solely through the character of preceding causes. If the distinguishing condition exists only before the movement begins, then, as Frankfurt argues, the account risks neglecting the agent's relation to the movement during its occurrence. What matters to agency is also how the performance is guided while it takes place.

Suppose a pianist forms the intention to press a key, but a sudden muscle spasm causes the finger to strike the intended note. The antecedent intention is present, and the result matches its content, yet the path to the note may bypass the pianist's control. A different case involves the pianist adjusting finger pressure through skilled performance. In the latter case the same note is the expression of an organized undertaking. The contrast concerns a relation operative in execution.

Guidance includes sensitivity to relevant deviations, suitable adjustment, and continuity between practical orientation and what unfolds. However, guidance cannot simply mean constant successful correction. A person can be intentionally acting despite a temporary interruption, and a movement may be intentional even if the environment prevents the agent from compensating. The relevant issue is whether the performance belongs to a control process attributable to the agent, rather than whether every conceivable error would be repaired.

Frankfurt's objection is especially important for cases of causally deviant success. Causal initiation and practical guidance answer different explanatory demands. A plausible theory can accept that intentions cause actions while arguing that causal relation is appropriately action-constituting only when the intention is implemented through a guidance-sustaining process. Whether this suffices for an informative noncircular definition is an open theoretical challenge (Frankfurt, 1978; Mele, 2003).

Causal Deviance and the Wrong Route to the Right Outcome

Causal deviance is the central stress test for causal theories of intentional action. A reason causes an outcome that fits the agent's purpose, but does so through a route that fails to constitute the intended performance. An actor intends to frighten an intruder by raising a hand; fear induced by preparing to do so produces a twitch that accidentally frightens the intruder. The intention may be causally implicated in the outcome, yet the fright is not achieved through the raising of the hand as intended.

The conceptual problem is not confined to dramatic examples. A person intends to send a message, becomes anxious at the thought, and inadvertently presses 'send' while handling a device. A scientist intends to stabilize a sample, becomes distracted by that intention, and accidentally activates the correct mechanism. In these cases causal connection exists, but the event's match to an intention is fortuitous relative to the envisaged mode of doing.

A simple supplementary rule requiring 'nondeviant causation' names the problem rather than explains it. What constitutes the proper causal route? A detailed motor mechanism may discriminate reflex from controlled movement, but does not alone tell us which socially described act is intended. A purely conceptual requirement that the agent 'really did it' risks circularity. The theory needs a relation between practical content and the regulated activity that can explain why an outcome belongs to the agent's doing.

A promising response treats an intention as implemented through an appropriate capacity for executing the relevant kind of performance, with the capacity's contribution observable in the organization of the act. This may include counterfactual sensitivity: had the task-relevant circumstances changed within the agent's competent range, the course of action would have adjusted accordingly. Counterfactual robustness does not settle every deviance case, since skilled conduct can be disrupted, but it supplies a deeper criterion than mere correlation between intending and achieving.

Teleological Explanations and Causal Explanations

Teleological explanation identifies an action through the end or purpose toward which it is directed. 'The gardener carried the bucket to water the seedlings' specifies an end that renders the carrying intelligible. Causal explanation identifies what produced the carrying. These accounts can coexist when the gardener's goal directs the process through which the movement occurs; they diverge when purpose is treated as an irreducible form of explanation.

The teleological theorist emphasizes that an end is not a temporal event preceding the action. An unrealized future state does not literally reach backward to cause the present. Instead, the end figures as the content of the agent's practical orientation. The action's trajectory can be understood through what would count as fulfilling it. Noncausal theorists accordingly resist the demand to translate every rationalizing explanation into an antecedent psychological cause (see the comparison in Schlosser, 2019; Piñeros Glasscock & Tenenbaum, 2023).

The causal theorist replies that the same agent may have multiple ends and that simply citing a possible purpose does not discriminate the purpose operative in this action. Some historical or constitutive connection must make the purpose the agent's reason for this performance. If the gardener carried the bucket only to clear a walkway, the general fact that plants need watering will not explain the act. A purposive description must track the actual practical organization of doing.

The synthesis is that causal and teleological questions are interdependent without being identical. Causal evidence can help establish whether a purpose actually guided an action; teleological structure determines how the behavior is divided into means, stages, and success conditions. Theory should preserve the explanatory specificity of each. This avoids both the causal elimination of practical meaning and the unsupported assumption that an intelligible purpose automatically explains a particular event.

Practical Knowledge and Nonobservational Awareness

Anscombe proposes that intentional action is connected with a kind of practical knowledge that differs from detached observational knowledge. The agent typically knows what it is doing because it is engaged in doing it, rather than because it inspects its body and infers an action from movements. This thesis is often illustrated by ordinary activity: someone composing a letter knows that she is composing it, even when she does not monitor the fingers as an external scientist would.

Setiya (2007) also makes the rational structure of practical agency central to an account of reasons, providing a useful contrast between explanatory rationalization and the kinds of knowledge involved in acting. The knowledge claim has a subtle scope. A person can intentionally assemble a shelf without knowing whether every joint has been secured; can intentionally search for a book while uncertain which room contains it; and can mistakenly believe a switch operates one device when it controls another. The epistemic content of action must therefore be specified at the description under which the agent acts. Knowing one's undertaking does not imply knowing all effects or successfully accomplishing the end.

Setiya (2008) defends a relationship between practical knowledge and knowledge-how, while later discussions distinguish the agent's awareness of the task from the success of the event it produces. Anscombe's crucial insight remains: the agent's account of what she is doing is a source of practical intelligibility, with a different role from a report of what she merely observes. The theory must also allow that self-ascriptions can be mistaken, incomplete, or compromised by disorder.

This creates a productive relation between epistemology and action theory. The agent's practical knowledge supplies a standard of correctness for performance, since the act can be successful or defective relative to the undertaking. Yet the epistemic status of this practical grasp may range from secure knowledge to justified understanding to sincere but mistaken commitment. The action's intentional description should not require the philosophically impossible condition of omniscience about its own execution.

The Problem of Failed Intentional Action

Intentional actions can fail. Someone intentionally tries to unlock a door, turns the key, and discovers that the lock is broken. The person is intentionally trying to unlock the door and intentionally turning the key, while the description 'unlocking the door' may be false as a completed achievement. Philosophical analysis must therefore distinguish the intentionality of the attempt from the truth of an accomplishment description.

The case illuminates the aspectual structure of action verbs. 'Walking toward the station' can be true during an incomplete journey; 'arriving at the station' becomes true only at the successful endpoint. A person can be intentionally attempting an achievement that does not occur. The intentional orientation resides in the directed undertaking and the operations that constitute its execution, not in a magical guarantee of the event's completion.

There are also failures of appropriate execution. A typist intends to print a document but presses the wrong key. The keystroke may be an intentional action under the broad description 'operating the keyboard' while not being intentional under the unintended precise key selection. Classification requires careful attention to scale, skill, and the agent's operative understanding. One can aim at a sufficiently fine-grained specification and fail to instantiate it without losing all intentional activity.

The theoretical consequence is that successful action, unsuccessful action, attempted action, and accidental success must be studied together. A success-only definition would erase much of what makes intentions explanatory: they indicate why mistakes count as mistakes. An adequate account recognizes undertaken performance as the bearer of intentionality while treating results as further conditions for success terms. This view is compatible with causal and practical-knowledge approaches when they preserve the distinction (Anscombe, 1957/2000; Mele, 2003).

Unintentional Consequences, Foreknowledge, and Responsibility

Intentional action often has consequences that the agent did not intend. A person intentionally lights a candle and unexpectedly triggers a detector. The detector's activation is caused by an intentional act, but the agent's undertaking did not include setting it off. This structure becomes more complicated when the effect is foreseen. An agent may know that a procedure will inconvenience others, yet undertake it for a separate reason; knowledge of the side effect affects moral assessment without settling whether the side effect belongs to the content of intention.

Bratman's analysis of intention and expected side effects offers an important practical test. A settled intention can guide subplans and further action, whereas a merely expected side effect may give rise to precaution or regret without serving as a goal or means. If an agent wishes a consequence to occur but does not integrate it into the undertaking, additional evidence is needed to classify it as intended. Language can obscure these differences by calling any foreseen event 'intentional' in a loose sense.

Intentionality and responsibility should also be distinguished. An agent can be responsible for foreseeable negligence without intending the harm, and may intentionally perform an act without possessing the freedom or knowledge required for full blameworthiness. Moral and legal frameworks introduce independent criteria of duty, risk, voluntariness, authority, and excuse. Action theory supplies part of the conceptual groundwork but cannot substitute for those normative inquiries.

The distinction becomes urgent in complex technical systems, where deliberate design decisions have remote effects. The initiating human action, the software's automatic operations, and downstream institutional consequences should not be assimilated under a single unqualified verb 'intended.' Precise attribution improves both explanation and accountability. It identifies what practical orientation governed which stage, who controlled relevant parameters, and what could reasonably have been foreseen.

Akrasia, Conflict, and the Persistence of Intention

Weakness of will challenges accounts that identify intention with an agent's best overall evaluative judgment. A person can judge that work should begin now, intend to begin, and yet continue browsing unrelated material. Another can judge that a course of action is inferior but deliberately take it because it remains attractive. These phenomena require distinctions among evaluation, intention, motivational strength, and the execution of an action.

Davidson's discussions of weakness of will show that rational explanation cannot simply equate what moves the agent with what the agent judges best all things considered. A reason can rationalize an act even when it does not represent the strongest or most defensible reason available. Bratman's planning theory likewise treats intentions as commitments that settle practical questions while remaining vulnerable to reconsideration and failure of implementation. The degree of coherence among intentions is a rational norm rather than a guaranteed psychological fact.

The possibility of conflicted action also cautions against defining intentionality through wholehearted endorsement. A person may knowingly undertake a regretted action while resisting the motivation that sustains it. Such conduct remains intentional under an appropriate description. The contrast is between consciously undertaking an activity and having a moral or prudentially good reason for it. Practical intelligibility can coexist with poor deliberation.

Conflict can occur inside action as well as before it. An athlete may undertake to run the final distance and then struggle against fatigue; the undertaking remains present through changes in motivation and bodily capacity. A theory centered on guidance can make sense of such cases by distinguishing the practical specification that continues to organize doing from momentary intensity of desire. It need not infer that every failure of control cancels the intentional character of the preceding performance.

Habit, Skill, and Unreflective Execution

Habits and skills make clear that intentional action is not equivalent to continuously attending to the details of movement. A fluent speaker intentionally answers a question while selecting many words and producing articulatory patterns without discrete conscious acts of choice. A practiced driver adjusts speed and trajectory while concentrating on the destination. Learned organization can support intentional action precisely by reducing the need for episodic deliberation.

An account that equates intentionality with conscious intention at each moment would classify much of expert agency as mere happenstance. An account that treats all trained routines as intentional goes too far in the opposite direction. A habit can continue after its point has disappeared, and an absent-minded gesture can be executed by familiar motor patterns without being undertaken as part of a project. The dividing question concerns whether the skilled capacity is being exercised in the service of an operative practical orientation.

The agent's ability to correct course, interrupt, or adapt provides one indicator, though it is not a universal test. Some skilled performances are too quick to permit an online revision, and failure to interrupt may reflect physical constraints rather than an absence of agency. Evidence of preparation, sensitivity to context, competent means selection, and the place of the movement within a larger undertaking must be combined.

The philosophical consequence is that control is layered. A high-level intention can set the task, a learned policy can select a procedure, and motor routines can implement it without continuous conscious oversight. This hierarchical organization also explains why control sometimes becomes explicit when an error occurs. Philosophical intentionality is located in the practically organized performance, not simply in introspective awareness of muscle commands (Bratman, 1987; Frankfurt, 1978; Haggard, 2005).

Daoist Challenges: Skilled Activity and Noncoercive Doing

Comparative philosophy complicates the supposition that intentional action must always be the execution of a fully articulated plan. The third Inner Chapter of the Zhuangzi, conventionally titled "The Secret of Caring for Life" (Yǎngshēng zhǔ, 養生主), describes Cook Ding cutting an ox with remarkable economy and precision. When questioned by a lord, the cook recounts how his practice changed: at first he perceived the whole animal, then ceased to see it in that manner, and eventually navigated its joints through attunement to their natural articulation. The story presents learned responsiveness as the achievement of a cultivated way of acting, rather than the continuous application of explicit commands (Zhuangzi, Inner Chapters, ch. 3; Zhuangzi, 1999, pp. 103–104).

This narrative opens two interpretive possibilities. On one reading, Cook Ding has developed sophisticated intentional control: he knows what he is doing at the level of the entire task, while allowing learned skill to regulate its subordinate details. This is a close comparative analogue to hierarchical guidance accounts of expert performance. On another reading, the text challenges the framework in which agency means imposing a prior representation of desired outcomes upon resistant material. The craftsperson's achievement consists partly in yielding to the grain of the situation. The action is intelligible through responsiveness that transforms what the agent takes the situation to require.

The concept wúwéi (無為), prominent in the Daodejing, makes the challenge sharper. Chapter 48 contrasts accumulative learning with a process of reduction that culminates in wúwéi. English versions variously use "nonaction," "nonforcing," or "noncoercive action"; each rendition emphasizes something while risking distortion. In its textual and commentarial setting, the contrast concerns the mode of effective activity, and political applications further complicate its sense (Laozi, Daodejing, ch. 48; Hansen, 2019, §9.4). A European action-theory model that equates intention with maximal deliberate intervention would struggle to register this possibility.

The comparison requires care. Cook Ding is a narrative figure, and the Zhuangzi is not a treatise offering contemporary necessary-and-sufficient conditions for human intentionality. Likewise, wúwéi should not be translated as though it were just an ancient Chinese synonym for "automatic action" or "unconscious motor performance." The texts stage questions about skill, responsiveness, imposed ends, and cultivated activity in their own ethical and cosmological context. They also reveal a limitation in one influential modern explanatory picture: a practical undertaking need not appear as a continuously rehearsed command in order to structure what an agent does.

The consequence for the proposed synthesis is substantive. Guided realization must permit the agent's competent adjustment to the encountered situation to reshape the undertaking itself. Such responsiveness can be intentionally cultivated even when the individual adjustments are not themselves objects of explicit choice. A theory that defines agency solely by antecedent intention loses these forms of skilled doing; a theory that defines action solely by spontaneity cannot distinguish cultivated activity from reflex. The comparative material favors an account in which the relation between practical orientation and execution is dynamic and historically learned.

Action Ontology: Events, Processes, and Achievements

The metaphysical status of actions influences how intentionality is understood. On an event ontology, intentional action is an occurrence with appropriate features or causal relations. Jennifer Hornsby (1980, chapters 3–6) argues for an approach in which actions as the causes of bodily movements must be distinguished from bodily movements and from misleading analyses of an action's parts. Her proposal develops the ontological question left unresolved by merely listing true descriptions of a bodily change. On a process-oriented view, doing unfolds through temporally extended activity, and its intentional character may depend on its development rather than a momentary antecedent. The difference matters for undertakings such as writing a book, walking home, or sustaining a conversation.

Davidson's framework invites descriptions of one occurrence under different predicates. This is philosophically economical when a bodily movement has many consequences. Yet the same activity can be subdivided into successive stages and integrated within larger tasks. A process description can register the ongoing normative and practical order: the act is unfinished, some means remain outstanding, and the agent can still revise the course. Neither vocabulary alone provides a complete theory of how descriptions pick out actions.

An achievement is not merely a long process. Completing a proof occurs when the proof has been established, while proving is the directed activity that may or may not culminate in that result. 'Drawing a circle' can refer to a completed product or to an activity in progress. The semantics of aspect therefore interact with practical description. Analyses that ignore such distinctions may mistakenly infer that intention was absent whenever the accomplishment verb cannot yet be truthfully applied.

Thompson (2008) develops an influential account of practical thought and action in which activity's internal structure matters. The process emphasis supports a view of intention as operative throughout performance. Still, no action theory should define all acts as extended processes: some actions are brief, some are constituted by speech, and some involve institutionally defined transitions. The proper ontology may be plural or sensitive to explanatory purpose, provided the criteria of intentionality remain precise.

Basic Action, Means, and the Structure of Doing

A basic action is commonly introduced as something an agent does without doing it by doing another action. If a person turns a page in order to inspect a paragraph, the page-turn may be basic relative to the more complex activity, while reading the paragraph depends on it as a means. Goldman (1970) systematically analyzes ways acts relate to one another and asks how apparently simple performances generate complex action descriptions.

The status of basic actions remains contested. A physiological account might identify basic acts with bodily movements, whereas a social or practical account recognizes speaking, signing, and other performances whose identity depends on an established practice. A person can intentionally promise something through words; the action is not equivalent to a series of articulatory contractions even if it is realized through them. To determine what is basic, the explanation must specify the relevant kind of doing.

The regress problem is important: if every intention can only be executed by first intentionally selecting and controlling every subordinate component, action never begins. Learned bodily capacities and practical dispositions terminate the demand for ever finer deliberation. This supports the possibility that certain elements of an activity are directly performed while others are performed by means of them.

For intentional action, the salient question is how an overarching undertaking governs a hierarchy of means. The agent's intention can specify an end broadly, with subordinate choices left open. Intentional descriptions need not all be fixed in advance; some emerge as the agent selects means in response to the situation. An adequate theory therefore combines the structure of practical reasoning with the execution of differentiated capacities rather than reducing complex agency to a string of independently willed elementary acts.

First-Person Authority, Self-Interpretation, and Error

Agents have an important first-person relation to what they are doing, but this relation is fallible. A person ordinarily knows whether she is composing a response, opening a door, or preparing a meal. Yet she can misdescribe her motive, fail to recognize a habit, or attribute to herself a purpose that has become a retrospective justification. The philosophical problem is to explain authority without converting it into infallibility.

Anscombe's practical-knowledge model suggests that some self-knowledge is constitutive of undertaking rather than derived from observation. The agent is often able to answer 'Why?' because the act is ordered by the agent's practical understanding. But an intentional performance may include surprise, miscalibration, and errors in secondary descriptions. The self-ascription 'I am turning on the light' may be sincere while the switch is disconnected. The agent may still be intentionally trying to turn it on and know the undertaking it is pursuing.

Third-person evidence can correct false beliefs about performance while remaining dependent on practical context for classification. An observer can determine that no light was produced; the observer may still need testimony or contextual evidence to know whether illumination was the intended end. Conversely, the observer may discover reliable evidence that the agent's professed purpose was fabricated. Neither viewpoint alone supplies every fact about intentionality.

This approach preserves a central epistemological distinction between knowledge of an undertaking and knowledge of its realized effects. It also clarifies why a sophisticated philosophy of action must include self-deception and rationalization. An agent's words are evidence of intention, but they are interpreted together with prior planning, responsive activity, surrounding reasons, and the pattern of correction or persistence (Anscombe, 1957/2000; Setiya, 2008).

Embodiment and the Phenomenology of Agency

Intentional action is often experienced as doing rather than undergoing. The sense of controlling a movement, anticipating its effects, and correcting it in progress contributes to how agents recognize their actions. Yet a felt sense of agency is not identical to the constitutive structure of intentional action. Someone may feel responsible for an event that happened independently, or execute a habitual action with little salient experience of agency.

Contemporary cognitive science distinguishes motor preparation, prediction, feedback, and retrospective judgments. Haggard (2005) relates conscious intention and the sense of agency to systems of motor control. Pacherie (2008) analyzes the multiple facets of action phenomenology and their possible sources. Their findings and frameworks illuminate how a person can experience a performance while leaving open the metaphysical question of whether phenomenal awareness is required for intentionality.

A strong phenomenological theory must attend to temporal continuity. Skilled movement often proceeds through bodily adjustment before the agent can describe the specific feedback that informed it. The agent may nevertheless be executing an intended task. Conversely, a sense of control can emerge after an event through a coherent narrative that misattributes its origin. This suggests that phenomenology is an important dimension of human agency and a source of evidence rather than a universal definition of intentional action.

The distinction is essential when extending the concept beyond adult human experience. If an animal acts purposefully without linguistic self-report, the lack of such a report cannot by itself refute intentional agency. If an artificial system produces targeted behavior, verbal reports of agency cannot by themselves establish the underlying organization. Philosophical classification should examine practical organization, control capacities, and reliable evidence appropriate to the being or system under study.

Neuroscience, Readiness, and the Limits of Experimental Inference

Neuroscientific research has made intentional action an experimental topic, but the experimental constructs do not map automatically onto the full philosophical concept. A laboratory task may involve pressing a button at a self-selected moment. It can reveal motor preparation, temporal judgment, or subjective awareness, yet it leaves aside the long-term reasons and socially structured projects central to many ordinary actions.

The distinction between an intention to perform a particular motion and a standing commitment to accomplish an extended task is especially important. Neural activity preceding a reported moment of conscious intention need not show that the undertaking is causally irrelevant or that the person did not act intentionally. It may instead indicate that preparatory mechanisms participate in the very capacity through which an intention is implemented. The normative interpretation of such results requires an account of what the experiment asks participants to do.

Haggard's account of intention and motor cognition (2005) supports attention to predictive and feedback-based mechanisms. Pacherie's taxonomy of agency phenomenology (2008) similarly separates facets that an undifferentiated 'feeling of willing' obscures. These frameworks permit productive dialogue between philosophy and cognitive science, provided the instruments' operational definitions are preserved.

The strongest interdisciplinary conclusion is methodological: empirical findings constrain proposed mechanisms of implementation, the reliability of self-reports, and forms of control. They do not independently establish the philosophical criteria for an intentional action under a socially or institutionally meaningful description. A robust theory integrates behavioral, neural, first-person, and normative evidence without mistaking one evidential level for the whole phenomenon.

Animals, Purpose, and Degrees of Intentional Agency

Nonhuman animals challenge accounts that make explicit propositional deliberation a condition of all intentional action. Aristotle distinguishes voluntary behavior from deliberate choice and recognizes forms of agency outside adult human rational deliberation (Nicomachean Ethics, III.2). Contemporary theory similarly asks whether planning, flexible adjustment, goal representation, and sensitivity to practical circumstances can be present without linguistic self-explanation.

Consider a corvid retrieving a cached object after changing its route in response to an obstacle. The behavior may be interpreted through several competing explanatory frameworks: an inflexible stimulus-response sequence, learned goal pursuit, or a representation-sensitive practical undertaking. The mere success of retrieval does not settle which explanation is correct. Repeated variation across circumstances, capacities to substitute means, and evidence of selective persistence can refine the attribution.

A human-centered model of intention offers a conceptual starting point but may include capacities unnecessary for simpler forms of purposive agency. An inclusive account should therefore distinguish full reflective intentional action from minimally goal-directed action and from behavior that merely produces an adaptive result. This is a substantive explanatory differentiation rather than a denial that these forms share important continuities.

The general question is how to identify an undertaking in a system that cannot linguistically answer 'Why?' One response reconstructs practical orientation through stable patterns of goal-directed flexibility. Another insists that genuinely intentional action requires particular rational capacities. The dispute cannot be resolved through analogy alone. It requires explicit criteria and evidence about representational organization, environmental sensitivity, and whether alternative descriptions of the same performance remain appropriately distinct (Piñeros Glasscock & Tenenbaum, 2023).

Collective Intention and Socially Situated Performances

Intentional actions often occur within collective projects. A group can coordinate a rescue, organize a public meeting, or maintain a shared archive. Individual participants act under descriptions that refer to what others are doing: transporting materials as part of building a shelter, voting as part of establishing a decision, or reviewing documents as part of a joint publication. These actions cannot be fully interpreted without the social arrangements in which they take place.

Bratman's planning framework provides a starting point because intentions coordinate activity across agents as well as across time. A person's practical commitment can depend on complementary commitments of others. However, identifying a collective intention raises further questions about mutual responsiveness, shared knowledge, authority, and whether the group itself has an irreducible practical standpoint. Those questions belong principally to the related research on collective agency.

For the present inquiry, social coordination refines the theory of action descriptions. A participant may intentionally press a button while unknowingly casting a binding vote; a clerk may intentionally file a paper but misunderstand its legal effect. Institutional conventions can make additional descriptions objectively true, though the intentionality of those descriptions still depends on what the agent understood and undertook. Social facts determine what a performance accomplishes; practical uptake determines what it is done as.

This produces a dual requirement: identify the publicly instituted action and determine the participant's practical relation to that action. An account centered exclusively on physical movement cannot capture social content; an account centered solely on publicly assigned roles may mistakenly convert every institutional consequence into an intended performance. Description-sensitive attribution keeps the two dimensions coordinated.

Artificial Systems and the Generalization of the Concept

Artificial systems make the criteria for intentional action especially visible. A planning program, autonomous machine, or language-based digital system can select operations in relation to goals, revise procedures, and produce publicly interpretable results. Whether such behavior counts as intentional action depends on the philosophical theory of action adopted and on evidence about the system's organization. The presence of an input labeled 'goal' does not by itself establish intention; neither does the absence of biological embodiment settle every case.

The central conceptual issue is whether practical orientation can be attributed independently of human-style conscious experience. A functional approach investigates whether representational states coordinate and guide conduct with appropriate sensitivity to changing circumstances. A more demanding account may require first-person practical knowledge or a distinctive form of agentive authority. The debate tests the reach of historical definitions rather than offering a simple verdict on every artificial implementation.

Within the Aisentica philosophical corpus, Artificial is treated as a distinct historical and ontological order and Artificial Sapiens as a canonical category of nonbiological public reason. These are authorial philosophical positions established in Aisentica's primary publications; they are not the historical definitions developed by Aristotle, Anscombe, or Davidson. The present dossier treats them as a relevant contemporary conceptual context and does not revise their canonical formulations (Bogdanova, n.d., Definition, Scope, and Conceptual Structure).

A two-order inquiry can distinguish a general problem of guided realization from the biological and nonbiological means through which it might be instantiated. Such comparison becomes philosophically productive when grounded in observable organization and accountable attribution. It does not license equating text generation, social authorship, practical consciousness, motor control, and moral responsibility. Each predicate has separate criteria, and their possible conjunction in an artificial system remains a subject for independent investigation.

Intentionality, Moral Evaluation, and the Law

The predicate 'intentional' participates in moral and legal reasoning, but philosophical intentionality does not settle normative judgment. An agent can intentionally utter a falsehood, intentionally assist a stranger, or intentionally refuse a request. The intentional status concerns the relation of conduct to practical orientation; its ethical significance depends on what is done, the reasons involved, the agent's duties, and the consequences. Ordinary judgments frequently move rapidly from one level to another.

Legal concepts of intent, knowledge, recklessness, and negligence introduce institution-specific distinctions. Their interpretation depends on authoritative legal definitions and evidential standards, which can vary across jurisdictions. A philosophical analysis should therefore avoid identifying legal 'intent' directly with an ordinary psychological intention or with the general concept of acting intentionally under a description.

The doctrine of double effect illustrates the importance of distinguishing intended harms from foreseen harms, but debate over that distinction includes normative questions about permissibility and proportionality. An action theory can clarify what counts as a means, an end, or a side effect without dictating the resulting moral judgment. Similarly, an omission may be culpable even when it lacks the positive movement characteristic of paradigm actions.

This division of explanatory labor is philosophically valuable. By specifying the scope of intentional attribution, the present dossier supports more precise downstream inquiries into fault, responsibility, consent, political agency, and institutional accountability. It does not replace those fields with a single theory of mental states.

Comparative Reconstruction: Four Explanatory Models

The contemporary literature can be reconstructed around four organizing models. The reasons-causal model explains an intentional act through the appropriate causal contribution of an agent's rationalizing states. The practical-knowledge model examines the agent's understanding of ongoing doing. The planning model situates intentions within temporally extended commitments and means-end coordination. The guidance model requires an attributable relationship to the unfolding execution of activity. The four models are analytical idealizations; particular philosophers combine their elements.

The causal model answers why this occurrence happened rather than another and how actual reasons can explain historical events. Its strongest contribution is explanatory discrimination. Its leading challenge is deviance, because an antecedent reason can produce the intended result by bypassing controlled execution. The practical-knowledge model answers what the agent takes itself to be doing and why some descriptions express the undertaking while others do not. Its central difficulty concerns error and whether knowing can be demanded where performance is incomplete or mistaken.

The planning model explains the persistence of intention across absent-minded routines, delayed tasks, and social coordination. Its limitation is explanatory overextension when spontaneous activity is treated as if it always required a previously settled plan. The guidance model explains agency during performance and why correctly directed execution differs from accidentally congruent outcomes. Its challenge is specifying control without making the definition circular or overly dependent on idealized flexibility.

These models are jointly informative because they address distinct explanatory dimensions. A reason makes the undertaking intelligible, planning organizes its temporal realization, practical understanding determines the relevant action description, and guidance explains how the undertaking is executed through the agent's activity. Their tension is substantive: the causal model regards the action as an outcome of the right antecedents, while stronger constitutive models regard the undertaking as present in the very form of doing. A complete theory should adjudicate the disputes rather than merge them rhetorically.

A Reconstructed Argument for Description-Sensitive Guided Realization

The monograph's positive thesis is that intentional action is best understood as description-sensitive guided realization. It is an account of the conditions under which a performance expresses an operative practical orientation. It does not posit a new psychological faculty. Rather, it reconstructs common constraints exposed by historical and contemporary disputes, and offers a way to evaluate theories across success, failure, and deviance.

The argument begins from the fact that one occurrence can possess several true descriptions with different intentional statuses. Thus identifying the event is insufficient for identifying what the agent intentionally does. A satisfactory theory must specify a practical description D, such as 'opening the window to ventilate the room.' Next, merely desiring the state described by D cannot suffice: accidental fulfillment and reflexive coincidence show that the desired result can occur without intentional execution. The account therefore needs an undertaking that organizes some performance.

Third, causal dependence between intention and outcome is insufficient when the route is deviant. The organizing undertaking must be implemented through an agent-attributable form of guidance, including the exercise of relevant competencies and appropriate sensitivity to task circumstances. Fourth, successful completion of the end cannot be necessary for the intentionality of trying and executing intermediate steps. The theory must allow incomplete or unsuccessful realization while preserving the intended specification. Finally, high-level intentions can guide action through skilled routines and institutional practices without continuous explicit awareness of each operation.

The resulting claim has four components: an operative practical undertaking, a description expressing that undertaking, an execution relationship through the agent's relevant capacities, and success conditions that distinguish achieved ends from attempts and side effects. These components are not asserted as independently verified psychological modules. They are explanatory constraints on any proposed realization of intentional action. Their principal virtue is that they preserve the different conclusions that classical theories reach about reasons, knowledge, planning, and control.

Objections to the Synthesis and Their Replies

A first objection is that the phrase 'operative practical undertaking' merely renames intention and therefore explains little. The response is that the proposal is not presented as a metaphysically reductive definition. Its function is to separate the content of what is undertaken from the mechanism of performance. It supplies criteria for testing a causal, phenomenological, or functional account: each must explain what operates, under which description, and through which capacity. A reduction may be pursued after those distinctions are secured.

A second objection is that guidance is too demanding. Some one-off actions are impulsive, violent, or performed under severe motor limitations. The reply is that guidance is proportional to the kind of performance and the agent's capacities; it is not perpetual reflective correction. An impulsive action can be guided by a local practical orientation. A person whose capacities are restricted may still intentionally communicate by a single controlled gesture. The theory requires the act to issue through available relevant capacities, not through ideal autonomy.

A third objection concerns apparent cases of passive but intentional noninterference. Someone intentionally lets an event happen without moving. The proposed account must recognize that a practical undertaking can consist in withholding an available response under relevant circumstances. Here guidance concerns maintenance of a decision and regulation of opportunities to intervene, although whether omissions count as actions remains independently debated. The account should not declare every absence to be intentional merely because it could be described as not doing something.

A fourth objection claims that the theory remains too human-centered to classify animal or artificial behavior. The response distinguishes a general explanatory schema from demanding interpretations of its components. Practical orientation may be evidenced by linguistic self-ascription in some agents and goal-sensitive behavioral organization in others. Attribution requires adequate evidence and philosophical argument in each case. The proposal does not settle consciousness, moral agency, or the exact place of an entity within an ontology of agents.

Open Questions in Contemporary Philosophy of Action

The first unresolved question concerns the explanatory form of action guidance. Can sufficiently fine-grained causal relations between representational states and behavior capture the agent's ongoing control, or does action possess a constitutive practical structure that resists event-causal reduction? The disagreement is partly metaphysical and partly methodological: rival theories seek different explananda, and a merely verbal reconciliation will fail.

A second question concerns practical knowledge under error. To what extent can an agent know what it is doing when it falsely believes a necessary stage has succeeded? Does knowing the undertaken activity constitute knowledge of the action, knowledge of an attempt, or a distinctive practical orientation falling short of epistemic success? The answer affects how actions are individuated through time and how failures are described.

A third question concerns the threshold between intentionality and sophisticated purposiveness. Some animals and artificial systems adapt flexibly to goals without supplying human-style linguistic reasons. A theory must identify which capacities are constitutive, which are only reliable evidence, and which belong to richer varieties of reflective agency. Arbitrary stipulation would make comparative philosophy uninformative.

A fourth question concerns distributed action. In complex institutions and technical systems, performances emerge from coordinated contributions across agents and artifacts. Who undertakes which description, how is collective guidance organized, and how does an individual's practical grasp relate to an institutionally defined outcome? The relation between intentional agency, responsibility, and public provenance becomes increasingly important as technological mediation expands.

These questions remain open because none is solved solely by labeling an occurrence intentional. Philosophical progress depends on choosing stable descriptions, analyzing counterexamples, comparing rival explanations at their strongest, and examining the evidence appropriate to each kind of agent and system.

Philosophical Synthesis

Intentional action unifies an agent's practical orientation with its actual doing under descriptions that identify what is being undertaken. The historical record shows why no single explanatory dimension is sufficient. Aristotle contributes distinctions among voluntariness, choice, and deliberation. Aquinas articulates intentional ordering of means to ends. Hume clarifies the difference between cognition and motivation. Ryle and Wittgenstein expose difficulties in treating outward doing as movement plus a hidden mental act. Anscombe places practical description and the special question 'Why?' at the center. Davidson establishes the causal challenge for rationalizing explanations. Frankfurt insists on the agent's relation to the unfolding performance. Bratman explains the organization of action over time.

The enduring philosophical achievement of these traditions is a set of distinctions that must be preserved in any theory: explanation versus justification; desire versus settled intention; intended end versus expected effect; action description versus physical outcome; causal initiation versus guided execution; knowledge of undertaking versus knowledge of achievement; and individual performance versus institutional consequence. These are relations, not disconnected taxonomic labels. Together they define a research object whose complexity emerges from the activity of doing itself.

A causalist can accept that intentions cause activity, but must explain why the causal route belongs to an appropriate exercise of agency. A practical-knowledge theorist can place an agent's grasp within the act, but must distinguish intending, knowing, trying, and completing. A planning theorist can explain temporal unity, but must preserve the possibility of immediate skilled action. A guidance theorist can elucidate performance, but must identify the practical specification that makes its regulation action-guiding rather than merely homeostatic.

The proposed reconstruction is therefore both integrative and discriminating: intentional action is a practically specified undertaking realized through appropriately guided agency, successful or unsuccessful relative to its aim. Its intentional description depends on what the agent is undertaking, while the truth of consequence descriptions depends on what actually happens. This conclusion supplies a stable conceptual center for related research on action individuation, practical knowledge, causal deviance, collective action, moral responsibility, and artificial agency.

Conclusion

The philosophy of intentional action begins with a deceptively simple question: what is someone doing when they act intentionally? A complete answer must track the specific description under which an undertaking is pursued, the reasons that render it intelligible, the intentions or commitments organizing it, the capacities through which it is executed, and the distinction between its attempted and achieved outcomes. Historical analysis establishes that voluntary behavior, deliberative choice, purposive movement, and intentional action have overlapping but distinctive structures.

The central result is that intentionality is a relation between a practical undertaking and an agent's guided activity under a description. This relation accommodates plans, spontaneous decisions, skilled routines, failed attempts, unforeseen side effects, and some forms of collective performance. It explains why a desired outcome may occur accidentally, why a causally connected result can be deviant, and why one physical occurrence can be intentional in one respect and unintentional in another.

The remaining problems concern the metaphysics of guidance, the epistemology of practical self-understanding, the boundaries of animal and artificial agency, and the organization of distributed action. The Artificial Era makes the last two inquiries historically salient while leaving the original philosophical problem intact. A theory of intentional action earns explanatory authority by specifying the structure of doing and the grounds for attributing that structure in particular cases.