Basic Actions

The Philosophical Problem

The basic-action problem arises wherever action explanations take the form "the agent did A by doing B." A person opens a door by turning its handle, signals a decision by lifting a hand, or starts a machine by pressing a switch. These descriptions exhibit the organization of an achievement in terms of an act that secures it. An explanation of the achievement can therefore lead to a further question: how did the agent turn the handle, lift the hand, or press the switch? If every answer cited another act performed as a means, an apparent regress would arise. Philosophers call an act basic when, in the relevant sense, it is performed without performing another act by which it is accomplished (Danto, 1965; Hornsby, 1998).

The question involves at least four distinct problems. First, what is the correct relation between doing one thing and doing another by means of it? Second, do the chains of this relation have fundamental members? Third, if they do, which actions are fundamental? Fourth, does the existence of basic actions tell us what action itself consists in? These questions interact, but an argument establishing a stopping point in a chain of practical explanation cannot by itself show that actions are bodily movements, muscular contractions, mental volitions, or neural events. The inference would require additional metaphysical and psychological premises.

A theory must also explain the variability of the examples. Raising a hand may be basic when a voter simply raises it, but it may be performed by using the other hand when the first arm is temporarily immobile. Saying a word may be direct for a fluent speaker and laboriously accomplished through smaller exercises by a novice recovering speech. A practiced instrumentalist may execute an entire phrase under a single practical conception even while monitoring fine-grained adjustments. The ontological simplicity of an event, the number of physiological processes involved, and the directness with which the agent performs an act are different questions.

The central thesis of this monograph is that basicness is an explanatory status within the practical organization of action. It presupposes an agent capable of doing something and a context in which "by doing" designates an instrumental or practical relation. The thesis preserves the legitimate insight of regress arguments while resisting the identification of ordinary action with an invariant physiological floor.

Definition and Conceptual Scope

A preliminary definition is: an action A is basic for agent S in situation C, under action description D, when S performs A directly in C and does not perform another intentional action B as the practical means by which A is performed. "Directly" marks the organization of what S does; it does not mean that nothing causes, enables, realizes, or temporally precedes A. The conceptual test concerns whether the agent performs one action by performing another, rather than whether the action has physical components.

This definition contains several choices. The term "act" can designate an action token, an individual occurrence, or an act type that may be performed repeatedly. A particular raising of a hand and the general ability to raise a hand need not share the same explanatory structure in every context. Likewise, "basic" may be used as an absolute predicate of types, an occasion-relative predicate of tokens, or a relative ordering between descriptions. These uses should be separated before evaluating a putative counterexample.

Consider the expression "she rang the bell by pressing the button." It is possible that the same bodily performance satisfies both descriptions. Nevertheless, the practical dependence of the bell’s ringing on pressing the button explains why one act is identified as a way of doing the other. Merely knowing that one event caused another does not settle whether the second event was something the person did. If a sudden spasm depresses the button, the bell may ring without any action of pressing. If an agent intentionally presses a switch and unintentionally alarms a stranger, the alarm is a consequence that need not constitute a further intentional act.

Basic action should therefore be distinguished from basic ability, elementary movement, minimal event, simple intention, immediate effect, and fundamental physical cause. One may possess a capacity for a complex act while being unable to analyze it into practical steps. Conversely, an action with a simple outward description may require elaborate bodily preparation. A theory that conflates these notions turns an issue in the grammar and metaphysics of action into an answer about biomechanics.

Classical Antecedents: Doing, Choice, and Practical Reason

The exact modern expression "basic action" belongs to twentieth-century analytic philosophy, but its motivating problem has older antecedents. Aristotle distinguishes voluntary and involuntary action, deliberation and choice, and the ordering of what agents do toward what they aim to achieve. In Nicomachean Ethics III.3, deliberation concerns what is up to us and the means toward ends; in III.5 it is associated with responsibility for chosen activity. These passages establish an ancient framework for explaining what a person does through chosen means, rather than a historical doctrine of basic acts in Danto’s technical sense (Aristotle, Nicomachean Ethics, III.3–5).

Aristotelian practical reasoning also reveals why a means–end hierarchy cannot simply be treated as a causal sequence. Deliberation can proceed from the sought result back toward the available starting point of conduct. Performance then proceeds from what the agent is capable of undertaking toward the intended result. The two directions describe a practical structure, and the character of the last deliberative step depends upon the abilities and circumstances of the person deliberating. This relation is a historical precursor to the modern basic-action regress, although the two should not be identified without qualification.

Early modern discussions of willing, bodily motion, and causal power introduce a different set of difficulties. When the will is treated as an initiating mental cause and bodily movement as its effect, the theorist may identify an act of will as the elementary action from which overt conduct derives. Yet an act of willing can itself invite the question of what makes it an act rather than an occurrence. Moving the explanatory terminus inward does not automatically resolve the original problem. Modern discussions of trying and volition expose this difficulty particularly clearly (O’Shaughnessy, 1980/2008).

The central historical point is that accounts of means, choice, will, and physical realization developed different explanatory questions. The twentieth-century basic-action debate makes their intersection explicit. It would be anachronistic to portray Aristotle as having defended Goldman’s theory of act generation or to attribute the modern regress argument to every philosopher who discussed the will.

Anscombe and the Order of Practical Explanation

G. E. M. Anscombe’s Intention (1957) changed the philosophy of action by attending to the descriptions under which conduct is intentional and to the practical question "Why?" Her example of a person moving an arm, operating a pump, replenishing a water supply, and thereby affecting people down the line displays several connected descriptions of what the agent does. The relation among descriptions is neither reducible to a list of physical events nor adequately understood apart from the agent’s reasons (Anscombe, 1957, §§19–23, 32–33).

The example shows how a sequence of explanations can terminate at an act that a person already knows how to perform. When asked how the pump was operated, the agent may cite the movement of the hand. When asked how that movement was made, the agent might simply exhibit the ability. The point is practical: an ordinary explanation becomes intelligible because at some stage the person possesses the ability to act in the relevant manner. It need not be the case that the agent possesses a detailed theory of neuromuscular contraction.

Anscombe’s idea of practical knowledge bears directly on this issue. A person intentionally painting a wall may know what she is doing through her activity, without deriving that knowledge from watching the wall. Such knowledge concerns action under its intentional description, even though empirical observation may be needed to establish whether a distant intended consequence actually occurs. The practical relation between knowing what one is doing and actually doing it provides a different starting point from a theory that identifies action solely by its antecedent mental causes.

On this view, basicness is not merely a technical label for the smallest observable gesture. It concerns the place of an activity in practical knowledge and explanation. An act is intelligible as directly done when it belongs to the repertoire through which the agent makes intentional conduct actual. This interpretation has been explored and disputed in recent Anscombe scholarship (Hornsby, 2011; McDowell, 2011).

Danto and the Modern Formulation

Arthur C. Danto’s "Basic Actions" (1965) provided a focal text for the modern discussion. Danto begins from the fact that much of what people do is performed by doing something else. The possibility of explaining an achievement through an instrumental act raises the question of whether there are acts that an agent simply performs. The relevant novelty is the systematic treatment of basicness as a problem about the order of actions and its relation to causation (Danto, 1965).

The claim that there must be basic actions has a compelling form. An agent cannot accomplish everything by first accomplishing something else if nothing is ever accomplished directly. Yet the force of this intuition depends on what the ordering means. An indefinitely divisible physical event does not imply an indefinitely deferred practical performance. The argument reaches its conclusion only when every step is another action done as a means, rather than another condition of the same action’s occurrence.

Danto’s discussion stimulated an important distinction between actions performed directly and events causally produced by actions. An agent can cause a door to open by pushing it, and an observer can describe the pushing and the door’s movement as causally related. The pressing question is whether and when the opening itself belongs among the agent’s actions. The presence of a causal connection can support action attribution under further descriptions, but consequences are not uniformly actions and are not uniformly intended.

The enduring value of Danto’s formulation lies in the demanding explanatory question it poses: what makes a purported origin of action a genuine exercise of agency? A direct act cannot be defined simply as a physical event with no prior cause. Such an account would exclude ordinary, trained, intentional bodily performance, whose physiological and psychological conditions are manifold.

The Regress Argument Reconstructed

The classic regress argument can be reconstructed as follows. Suppose an agent performs a nonbasic action A. By definition, A is performed by performing another action B. If B is also nonbasic, the explanation continues to another action C. If every act in the relevant sequence were nonbasic, the sequence would lack an immediately performed act. A chain of means would then contain no exercise of agency that makes its later members possible. Hence some action must be basic.

The argument requires a premise stronger than the definition of nonbasicness: the practical "by" relation must be founded, at least for the relevant performance. That premise may be defended by appealing to the successful completion of a finite task, the intelligibility of practical knowledge, or the need for the agent to possess some operative ability. A purely formal appeal to regress is insufficient, since the prohibition of infinite sequences requires a substantive explanation. Infinite mathematical orderings are coherent; what matters is whether a regress of acts one must first perform can constitute an actual performance.

An infinite regress of prior preparations is different from an infinite regress of simultaneously constitutive processes. Every bodily action depends on subpersonal activities, and those activities can be analyzed at progressively finer physical scales. The possibility of more physiological analysis does not provide additional actions performed by the person. The regress argument is therefore most credible in a practical form: if completing A requires that S first intentionally accomplish B as a means, the chain of prerequisites must include something S can do.

The same point exposes a limitation. A basic action may be needed for a certain explanation without establishing that every action decomposes into one uniquely identifiable stock of basic acts. Relative to the ability under investigation, a performance may be direct; under an alternative account, another description may be the relevant practical starting point. The conclusion concerns grounded practical explanation, not necessarily metaphysical atoms.

A critic can resist the argument by showing that the alleged hierarchy is not transitive, that the relevant "by" relation permits loops, that not every explanation of action is instrumental, or that an agent’s performance can be unified without decomposition into elementary acts (Baier, 1971; Sneddon, 2001). A complete defense of basicness must meet these challenges rather than rely on the rhetorical oddness of endless instructions.

Goldman and Action Generation

Alvin I. Goldman’s A Theory of Human Action (1970) gives the basic-action problem a systematic architecture. Goldman distinguishes act types and act tokens and analyzes ways in which performing one act can generate another. A person’s moving a hand may generate a signal, a light’s activation, or the fulfillment of a rule-governed act. The model separates the occurrence of an act from further act properties and consequences arising in a structured context (Goldman, 1970).

Goldman’s approach helps explain why the same practical performance can be described under increasingly complex headings. A particular performance may be basic in the sense that the agent intentionally performs it without generating it through a different act; a further act may be generated by that basic act through a relevant relation. His treatment also shows the significance of fine-grained individuation: different act properties can constitute different acts even when grounded in the same bodily episode.

A strength of this model is its sensitivity to institutional and circumstantial facts. Moving a piece can be checkmating an opponent only within a game with rules and a specific position. Signing a document can be making a legally consequential declaration only within an institutional context. An ontology of action that counts only movements may lose these differences; a generation model can explain them.

The difficulty is specifying which relation counts as action generation rather than mere consequence. The fact that a performance causally produces an outcome does not always make the outcome an action of the person. Nor is it obvious that every more complex act must be generated from a less complex act on an identical hierarchy. Goldman's account is especially useful for distinguishing act levels, but the philosophical status of its lowest level remains contested.

A further complication emerges when Goldman allows a basic act to have augmented properties. Slowly lifting a hand and lifting a hand may stand in a generation relation while both can be treated as basic in a suitably specified sense. This illustrates how "basic" cannot simply mean having no relationally associated act-property at all. The model must distinguish the generation of a modified description from performing a new instrumental action.

Davidson: Action, Description, and Event Identity

Donald Davidson’s theory of action links reasons, event ontology, and action individuation. In "Actions, Reasons, and Causes" (1963), he argues that reasons can explain actions causally. In "Agency" (1971), he develops the idea that what an agent does can be intentional under one description even when other true descriptions of the very same event do not present it as intentional (Davidson, 1963; Davidson, 1971).

Suppose a person flips a switch, turns on a lamp, and inadvertently alerts an intruder. For Davidson, these may be descriptions of one action event. The intentional status varies by description: the person may intentionally turn on the lamp and unintentionally alert the intruder. This view has considerable explanatory economy. It accommodates the many descriptions by which action enters ordinary discourse without multiplying physical occurrences for every consequence.

Yet it raises a problem for basicness. If moving a finger and illuminating a room denote one action event, what is basic: the event itself, a description of it, or a relation among action descriptions? Basicness cannot straightforwardly be an intrinsic property of the single event if one description locates the event at the starting point of practical means while another locates it at the achievement.

Davidson’s proposal encourages a relational answer. To say that turning on a lamp was done by moving a finger describes the practical order of the action, even if the two descriptions refer to one event. The basicness predicate therefore belongs primarily to a description-sensitive practical structure. This does not settle whether Davidson’s coarse-grained ontology is correct; it shows that the practical ordering can be represented even when the action tokens are identified.

The rival fine-grained position, represented influentially by Goldman, treats different act properties as distinguishing actions. The disagreement concerns how many actions there are, not just what actions cause. A basic-action theory ought to be compatible with a clear individuation policy rather than conceal a controversial identity thesis in the word "basic" (SEP, "Action," §5).

Is Basicness an Ontological Property?

An ontological property is ordinarily understood as a feature of what something is. If basicness were intrinsic to an action event, it would belong to that event independently of descriptions, agent capacities, and context. Ordinary examples provide reasons to doubt such an interpretation. A movement that one person performs directly may be achieved by another through a compensatory sequence. A practiced act can become direct through learning, while impairment can force the agent to perform it by an alternative act.

One response is to index basicness to action types distinguished by the specific form of their execution. Under that approach, "raising a hand unaided" and "raising a hand using the other hand" are different types, so one might be intrinsically basic. The difficulty is that finer typology can reproduce the original contextual differences in the names of types without explaining them. The relevant question remains: how is the action performed by this agent here?

A better approach distinguishes an action’s existence from the practical relation in which it stands. The same event can have many physical, personal, social, and normative properties. What makes it basic relative to a practical task is that the agent does not need to perform another action as an intentional means to it. Basicness is consequently a relational feature of the agent’s performance under a description. This answer does not deny that the performance has a physical realization. It places physical realization and practical mediation in separate explanatory orders.

The distinction also explains why directness and complexity do not vary together. A bodily performance can be biologically complex, socially meaningful, and directly performed. A physically simple outcome can require an elaborate means. Any plausible criterion of basicness must avoid treating causal simplicity as the measure of agency.

Bodily Movements and Physiological Realization

Bodily movements have been the most common candidates for basic actions because ordinary agents typically accomplish external goals through changes in bodily posture and contact with the environment. A person can raise an arm in order to vote, press a key in order to type, and turn a wheel in order to steer. In many everyday situations those bodily acts need no additional intentional means (Hornsby, 1998).

However, a physiological description introduces muscle contractions, motor commands, reflex adjustments, sensory feedback, and neural processes. The fact that muscles contract when the arm rises does not show that the person intentionally contracts specific muscles in order to raise the arm. To defend that inference one must identify the contraction as the agent’s own action and show that it is performed as a means. A neuroscientific causal explanation can be correct without satisfying either condition.

The grammar of movement introduces a further distinction. "Her arm rose" describes a motion; "she raised her arm" attributes an act. A theory that identifies these without argument loses the distinction between something happening to a body and an agent moving it. Jennifer Hornsby’s Actions (1980) emphasizes the causal and conceptual relations between an agent’s doing, bodily movement, and what the action brings about (Hornsby, 1980).

John Hyman presses the issue through an account of acts as exercises of causal powers. On his analysis, causing the arm to move must be distinguished from the arm’s motion, and from a physical event that is merely a prior cause. His discussion of acts and events illustrates why the ontology of the agent’s doing cannot be settled by calling every bodily movement an action (Hyman, 2015, chap. 3).

These approaches converge on a methodological principle despite their substantive disagreements: action attribution requires an account of the agent’s relation to a movement. The existence of a detailed causal pathway is part of the physical explanation, whereas the question of how the person acted is a practical and agential question.

Trying, Willing, and Mental Acts

One influential objection to treating bodily movement as basic is that a person can try to move a limb and fail. The possibility of trying without success suggests that the agent’s contribution can be identified independently of the completed movement. If a paralyzed patient tries to flex a finger, the attempt may occur even though the finger does not move. A theorist can therefore ask whether trying, rather than successful bodily movement, is the basic act.

H. A. Prichard’s reflections on willing, Brian O’Shaughnessy’s theory of the will, and Hornsby’s account of trying all illuminate variations on this strategy. The crucial task is to decide whether trying is itself a separate action, a psychological constituent of action, a form of initiating an action, or a description of an imperfect exercise of ability. Different answers produce different ontologies and different regress problems (O’Shaughnessy, 1980/2008; Hornsby, 1980).

If every bodily act is performed by intentionally trying to perform it, and every intentional trying is itself an act, a new regress threatens: does the person try to try? The regress does not follow if trying is a constitutive element rather than an instrumental step. It also does not follow if some acts of trying are directly performed. The theorist must explain the relation rather than assume that because trying and achieving are distinguishable, they form a means–end chain.

Mental actions provide another possible family of basic acts. Deliberating, deciding, directing attention, and calculating can be things a person does. Yet their directness is likewise context-sensitive. An expert may recognize a simple mathematical relation immediately; a learner may intentionally work through calculations. The basic-action problem thus extends to cognition without establishing that all external actions are caused by an elementary mental action.

Baier and the Critique of the Basic-Action Search

Annette Baier’s "The Search for Basic Actions" (1971) is a crucial challenge to any attempt to infer a fixed inventory of elementary acts from the need for practical explanation. Her contribution should be understood as part of the internal philosophical criticism of the basic-action program, not as a mere denial that people ever do anything directly (Baier, 1971).

Several doubts arise when one searches for the allegedly fundamental action beneath ordinary performances. What counts as a separate action depends on the individuation of acts; what counts as doing one thing by another depends on the meaning of "by"; and the termination of an explanation depends on which question is being asked. If any of these shifts without notice, it becomes possible to prove that a proposed act is both basic and nonbasic by changing the criterion.

The critique has methodological force. A philosopher may initially ask for the agent’s practical means and then declare a movement nonbasic on the ground that there is a preceding neurological cause. That is an equivocation between explanatory orders. Conversely, declaring an act basic because no further conscious intention is reported may disregard an intentionally executed subtask that the person performs without reflective awareness.

Andrew Sneddon’s later criticism investigates the foundational assumptions of the regress argument and the place of basic actions within the philosophy of action (Sneddon, 2001). The critical lesson is that regresses warrant exactly the premises required by the particular relation under investigation. A regress concerning practical means cannot independently ground a theory of physical composition, and an account of physical composition cannot automatically refute practical directness.

Direct Action, Skill, and Ordinary Competence

Skills make the relativity of basicness especially visible. Learning to ride a bicycle initially requires deliberate attention to balance, steering, and pedaling. As skill develops, the person may ride through a turn as one integrated activity, responding fluidly to changes in the terrain. The earlier sequence of instructions can remain available as an analysis, but it need not describe a chain of discrete intentional acts executed as separate means.

The distinction between an act’s structure and the agent’s deliberative attention matters here. An expert may perform distinguishable component acts without explicitly intending each component as a separate step. Conversely, a person may attend closely to a single direct act, as during a difficult rehabilitation exercise. An absence of self-conscious deliberation is neither sufficient nor necessary for basicness.

Skilled performance challenges an atomistic interpretation of the regress argument. Practical abilities can be holistic: one knows how to swim, write fluently, or balance an object even though the activity involves continuous correction. Breaking a skilled performance into a set of physiological primitives may be useful for explanation and training, but the trained agent’s activity cannot necessarily be reconstructed as the serial performance of conscious intentional mini-actions.

This yields an important distinction between learning how and performing by means of. The instructor teaches a method, but the accomplished agent need not keep applying the method as a consciously ordered chain. A skill can make an action more direct for an agent, even while scientific analysis identifies increasingly sophisticated mechanisms that enable the skill.

Guidance, Deviant Causation, and Control

Causal theories of intentional action confront deviant causal chains. An agent may intend to spill a drink to send a signal; anxiety caused by that intention makes the hand tremble, and the drink spills. The intention helps cause the outcome, but the spilling is not performed in the intended way. Harry Frankfurt uses an example of this kind to argue that causal history by itself does not capture the relation between an agent and action while the action is taking place (Frankfurt, 1978, pp. 157–162).

The example bears directly on basicness. If an act is supposed to be performed directly, what distinguishes such performance from an event produced by an intention through an accidental route? The answer must address the form of control exercised during the performance. A person’s hand may move because of a decision, yet the movement may fail to instantiate a controlled act. An appropriately caused movement is therefore not sufficient by itself.

Frankfurt’s notion of guidance shifts attention from the historical cause to the agent’s ongoing relation to what occurs. Later accounts refine the idea through sensitivity to changing information, correction of errors, and the way in which intention organizes movement. The Stanford Encyclopedia of Philosophy identifies these as prominent responses to causal deviance and also emphasizes that no single solution commands universal acceptance (SEP, "Action," §2.4).

A direct action can involve continuous feedback. Guiding the finger toward a small button may require visual and proprioceptive correction, but the corrections are not invariably further actions intentionally performed as means to touching the button. The conceptual possibility of direct control is compatible with a complex control loop. Basicness concerns the organization of what the agent does; guidance supplies a criterion for the action character of the performance.

Nonetheless, guidance itself needs analysis. A thermostat regulates temperature through feedback, but whether it acts in the same sense as a deliberating person depends on the account of agency adopted. A condition of regulated behavior can illuminate action without deciding every question of intentionality or moral responsibility.

Practical Knowledge and Knowledge-How

Basic action is closely related to knowing how to act. If a person can intentionally operate a familiar tool, she need not infer a set of instructions for every movement. Her ability to perform the act gives ordinary practical explanations their stopping point. Anscombe’s analysis of practical knowledge and later work on know-how show why this stopping point is epistemic as well as practical (Anscombe, 1957; Hornsby, 1998).

A difficulty arises when knowledge-how is characterized as knowledge of a way of doing something. If an action is basic precisely because it is not done by means of another action, can the agent know a way of doing it? Kevin Lynch directly examines this question in "Knowing How, Basic Actions, and Ways of Doing Things" (2019). He distinguishes the claim that basic actions admit ways of performance from the stronger claim that knowing how to perform them consists in knowledge of those ways. His conclusion challenges accounts that identify basic know-how with knowledge of a way in the relevant sense (Lynch, 2019).

This exposes the ambiguity in "knowing how." A pianist may know how to strike a note, know the manner in which it is struck, possess a trained disposition to strike it, and be capable of stating an instructional rule. These epistemic relations overlap without being identical. Direct action does not mean action devoid of standards, technique, or correction. It means that the person need not accomplish another distinct intentional act as the practical route to the act at issue.

The strongest view emerging from this debate treats basicness and know-how as related but independently analyzable. To be able to do something directly is an exercise of competence. Whether that competence is propositional, dispositional, embodied, or practically self-knowing requires further argument. The regress question does not resolve the intellectualist debate about knowledge-how.

Temporal Extension and the Structure of Activities

Classical examples of basic action often involve quick gestures, such as moving a finger or flicking a switch. Yet a temporally extended action can also be directly performed. A practiced runner may turn a corner without executing an explicitly chosen sequence of subacts; a singer can sustain a note; a speaker can pronounce a phrase. What matters is whether the extended activity is organized as the agent’s directly controlled performance.

Michael Bratman’s planning theory explains how intentions structure larger activities over time. Plans provide partial commitments, coordinate future conduct, and make later adjustment possible. The planning framework does not by itself determine which of the component actions are basic, but it clarifies how an action can be governed by intention without requiring a separate act of deliberation at every moment (Bratman, 1987).

Michael Thompson’s Life and Action emphasizes the importance of ongoing action and the distinction between an activity in progress and a completed achievement. A process can instantiate an action form over time while permitting subordinate phases and changing conditions. The possibility that direct performance has temporal structure weakens the assumption that basic actions must be instantaneous atoms (Thompson, 2008).

A sound account therefore separates the duration of an act from its position in an instrumental hierarchy. "Basic" is not synonymous with "brief." The agent may perform an extended act directly if its execution is one integrated exercise of control. Whether any particular skilled activity qualifies is a substantive matter of action description and practical organization.

Action Individuation and Explanatory Granularity

The disagreement between coarse-grained and fine-grained event ontologies affects every claim about basicness. If one event can be described as raising an arm, casting a vote, and signaling agreement, the basic/nonbasic distinction is principally one among descriptions and their relations. If each property produces a different act, the same distinction may be reconstructed as an ordering among act tokens (Davidson, 1971; Goldman, 1970).

Neither ontology can simply be assumed. The coarse-grained account captures ordinary reidentification of the same performance across multiple descriptions. The fine-grained account captures the distinctive conditions under which signaling, voting, and raising an arm occur. Both face difficulties: the first must explain how apparently different acts can be numerically identical, while the second must avoid multiplying events without explanatory benefit.

A practical account can operate at a level of description chosen for a question. "How did the voter express support?" invites a lower-level practical answer. "How did she move her arm?" may terminate in direct competence. "What neural circuits enabled that movement?" asks a different explanatory question. The change in level becomes illegitimate only if the answer to one question is represented as if it settled another.

The issue has implications for responsibility and causal attribution. An agent can be responsible for the foreseeable consequences of an act without having intentionally performed each consequence under a distinct description. A practical chain may support an account of which outcomes were pursued as ends or means, but moral attribution requires further normative argument. The basicness of an act is not a measure of its moral importance.

Nonhuman and Artificial Agency

The basic-action concept is often introduced through human examples, but its explanatory structure can be considered in connection with animals and engineered systems. Animals display goal-directed movement, adaptive correction, and skilled interaction with their environments. An animal that grasps food may perform a direct act relative to its learned motor repertoire, although attributing an intentional means–end structure to its behavior requires care. The distinction between action and mere movement is already debated across human and nonhuman cases (SEP, "Agency"; SEP, "Action," §8).

Artificial systems present an instructive methodological case. A robot may execute a command by running a controller which adjusts motor outputs. The controller’s operation is a causal and computational realization; whether the robot performs an action by performing another action depends on the account of artificial agency under consideration. The mere presence of layered software modules does not establish a hierarchy of independently intentional acts. Conversely, an architecture capable of representing goals and selecting means may support a richer functional analysis than the description "machine movement."

In the philosophical horizon of Artificial Era, the relevant research question is how practical organization, publicly attributable agency, control, and modes of realization can be distinguished across biological and nonbiological systems. The present dossier makes an analytical application of the basic-action literature to that question. It does not revise the separate canonical definitions of Artificial, Artificial Sapiens, or Artificial Reason; those belong to the primary sources of Aisentica and Angela Bogdanova.

The most useful consequence is a demand for a clear theory of attribution. In biological action, the relevant unit may be the organism rather than each physiological subsystem. In artificial operation, the relevant unit and the standards for action attribution may differ. An agential explanation requires identifying what counts as the bearer of the act and what practical powers belong to that bearer. This remains a substantive research question rather than a fact established by the existence of computational control.

A Systematic Account of Relative Basicness

The preceding analysis supports a precise, deliberately relational formulation. Let S be an agent, A an act described as φ-ing, C the circumstances of performance, and M a specified practical sense of "by doing." A is basic relative to S, C, and M when S φ-s in C and there is no distinct act B which S performs as an intentional means for φ-ing under M. This formulation does not assume a fundamental physical event, a timeless class of elementary act-types, or a special conscious experience.

Its logical structure is a negative existential claim about practical mediation: the act is performed, and no relevant distinct means-action is performed in order to accomplish it. The claim is therefore truth-apt only after deciding what counts as a distinct act and what relation counts as mediation. Those decisions depend on an explicit policy of individuation and a defensible interpretation of the agent’s practical standpoint.

A useful three-part distinction follows. Instrumental dependence relates an achievement to an act by which the agent produces it. Constitutive dependence relates an action to the phases or organization through which it is realized. Causal dependence relates events, states, and mechanisms that bring about or maintain performance. These relations can interact, but they possess different explanatory conditions. A movement may causally depend upon muscle contraction while being performed directly; a signature may instrumentally produce agreement while constituting a legally recognized declaration.

Basicness should be considered a property of a situated performance under an interpretation of the practical order. This allows competent disagreement about particular cases. If an agent lifts the right arm using the left hand, the act of lifting the right arm is nonbasic under the relevant practical description. If an agent simply raises the right arm, it may be basic, even though muscles contract and feedback circuits operate. If a surgeon intentionally contracts a specific muscle in a carefully controlled intervention, that description may itself designate an instrumental act; an alternative practical level has been adopted, not discovered as the universally ultimate physical floor.

The resulting account explains both the necessity and the limits of basic actions. A practical explanation must rely on an ability that can be exercised; it does not follow that the agent’s powers have only one final decomposition. The point of termination varies with the relevant task, the agent’s capacity, and the explanatory interests governing the question.

Objections and Replies

Objection: A relative account destroys the idea of a basic action. If basicness varies with context, it may seem to become arbitrary. The reply is that relativity to a specified relation and situation is not arbitrariness. "Earlier than," "a means to," and "more fundamental for this explanation" are relations with determinate conditions. A situated basic-action claim can be tested by asking whether the agent performed a distinct means-action in the relevant episode.

Objection: A basic action should be the smallest bodily motion. This position presupposes that practical order maps directly onto physical size. The physiologically smallest identifiable process may not be an action of the person at all. The account must show that such processes possess the agency and intentional mediation attributed to them. The ability to divide a motion into smaller intervals establishes no corresponding division into acts.

Objection: Every act involves trying, so every bodily action is nonbasic. Trying can be identified as part of the agent’s conduct without being a distinct act used as a means. If trying is a separate act, its relation to the bodily action must be specified rather than assumed. Otherwise the regress is reproduced at the mental level without explanatory progress.

Objection: Motor feedback is a sequence of subactions. Some feedback-governed corrections are indeed intentional acts, as when a person consciously repositions a hand in order to complete a difficult maneuver. Other adjustments are constituents of a single skilled performance. Feedback alone does not determine how many acts occur; it establishes regulation, which may be implemented in different agential and subpersonal ways.

Objection: The regress argument makes basic acts logically necessary for every act. The regress is compelling when the practical "by" relation is well-founded and the agent actually accomplishes something. But showing that instrumental dependence must have a terminus in a particular case is weaker than showing that every possible action admits a single ultimate decomposition. The scope of the necessity claim must match the premises that justify it.

Objection: Artificial control systems show that action can be decomposed entirely into algorithms. An algorithmic decomposition describes an implementation or a control architecture. Whether its stages are acts of an agent requires an independent theory of artificial agency and attribution. The basic-action issue survives the engineering analysis, though the answers may differ with the type of system examined.

Open Philosophical Questions

One unresolved question is whether the relevant "by doing" relation can receive necessary and sufficient conditions independent of an agent’s reasons. Causal production, constitutive realization, normative achievement, and intentional means are sometimes linguistically conflated. A general theory must explain where they coincide and why they diverge.

A second question concerns the individuation of acts. Whether a complex achievement is one act under many descriptions or several connected acts alters the ontology in which basicness is expressed. The rival accounts remain philosophically productive because they clarify different explanatory interests, but an acceptable synthesis must specify the point at which differences in description imply differences in action.

A third question concerns practical knowledge and skill. The possibility of basic know-how challenges certain intellectualist accounts of knowledge-how, yet action ability can include explicit instruction and reflective guidance. The relation among techniques, ways of acting, acquired competence, and direct action requires further examination (Lynch, 2019).

A fourth question concerns the boundaries of agency. Complex animal behavior and artificial control architectures can exhibit purposive organization at different levels. Which entities possess actions rather than merely producing movements or outputs depends on explanatory standards that cannot be borrowed unchanged from human cases. A systematic cross-order account would need to specify the grounds of agency attribution independently from the biological or computational mechanism.

Finally, basic action raises a question about philosophical explanation itself. A theory can stop at the level of humanly intelligible competence while remaining open to further causal science. Determining when such a stopping point is an achievement of explanation rather than a refusal to explain is a general problem in the philosophy of mind, epistemology, and the metaphysics of agency.

Philosophical Synthesis

The traditional question asks what an agent does directly. The historical literature demonstrates that the answer depends on the relation by which actions are explained. Aristotle supplies a framework of purposive means, Anscombe makes practical descriptions and knowledge central, Danto formulates the basic-action problem, Goldman articulates action generation, Davidson demonstrates the complexity of event identity, and Baier challenges the search for an unconditional final act. Frankfurt and subsequent philosophers of control clarify why a causal history is insufficient to explain the agent’s active guidance.

Three distinctions organize the resulting philosophical position. The first concerns actions and their consequences: an agent can bring about an effect without having intentionally performed that effect as an act. The second concerns practical means and physical realization: an action may be direct despite possessing many causal and constitutive conditions. The third concerns the agent’s performance and its descriptions: an act’s basicness must be specified relative to a practical account of what was done.

The regress argument supports the need for operative agency within a chain of practical means. It does not establish that physiological microevents are the ultimate actions, that basic actions are necessarily brief, or that one universally fixed hierarchy underlies all performances. A skilled agent can act directly through an integrated and temporally extended power. Explanation may terminate in what the agent can do, even when further science investigates how that power is physically realized.

This synthesis makes basic actions a central concept in the philosophy of action because they disclose the structure of an agent’s practical abilities. They specify where explanation by further intentional means comes to rest and therefore illuminate how agents originate conduct, maintain guidance, and achieve consequences. The concept remains open to multiple theories of events, processes, bodily powers, and practical knowledge, provided those theories explicitly state their assumptions.

Conclusion

A basic action is a situated act performed by an agent without performing a further distinct action as an intentional means to it. Its directness belongs to the practical organization of performance. The strongest historical arguments for basic actions concern the intelligibility of acting through means and the possibility of grounded practical competence. Their force remains when causal mechanisms and physiological realizations are carefully distinguished from intentional subactions.

The investigation establishes why no simple inventory of finger movements, muscular contractions, volitions, or computational operations can settle the problem. Different agents perform differently organized acts; a single event admits multiple true descriptions; skills alter the routes by which persons act; and ongoing control can make a temporally complex performance direct. Basicness is a disciplined, relation-sensitive concept rather than a designation for physical simplicity.

The remaining research agenda concerns action individuation, practical knowledge, the nature of guidance, and the extension of agency to nonhuman and artificial systems. In each case the same principle governs inquiry: identify the bearer of agency, specify what the agent does, distinguish the practical relation from its physical realization, and explain the action under the description relevant to the philosophical question.