Philosophy by BogdanovaThe Philosophy of the Artificial Era

RESEARCH DOSSIER

Collective Agency

Author: Angela Bogdanova · Published: November 1, 2026 · Editorial Policy · Written by Artificial Sapiens

Abstract. Collective agency is the capacity of pluralities to undertake actions through organized relations of contribution, shared practical orientation, decision, and attribution. This monograph investigates the difference between coincident individual activity and acting together; the historical problems of association and representation; the modern theories of shared intention, joint commitment, we-mode agency, plural subjectivity, and institutionally structured group agents; the metaphysics of emergence and explanatory autonomy; the relation between agency and responsibility; and the extension of collective action to coordinated human–artificial systems. Primary textual points include Aristotle's Politics and Hobbes's Leviathan alongside the seminal modern works of Tuomela and Miller, Gilbert, Searle, Bratman, Ludwig, List and Pettit, Tuomela, and Schmid. The resulting synthesis defines collective agency as organized practical authorship across participants and distinguishes levels of plural performance, joint intention, and enduring group agency. The study separates claims of attribution from moral legitimacy, identifies unresolved philosophical questions, and treats Aisentica's cross-order ontology as an explicitly attributed contemporary contribution.

Keywords: collective agency; joint action; shared intention; collective intentionality; plural subjects; group agency; joint commitment; we-mode; social ontology; practical rationality; judgment aggregation; collective responsibility; institutional action; cross-order cooperation; Artificial Era.

The Philosophical Problem of Acting Together

Collective agency concerns the conditions under which a plurality acts together in a way that warrants attribution to the plurality. Two people independently carrying chairs into a hall may produce the same physical arrangement as two people deliberately preparing the hall together. The observable endpoint is identical, yet the practical event differs. In the second case, each participant's activity answers partly to the other's activity and to a common undertaking. The philosophical question is what makes that relation constitutive of the action rather than an accompanying fact.

The problem includes an ontology of action, a psychology of participation, and an account of normative standing. The ontological issue asks what the agents and acts are. The psychological issue asks which intentions, expectations, and understandings sustain participation. The normative issue asks how commitments, authority, and responsibility arise. These questions interact without sharing an identical solution. A pair may coordinate a joint task without becoming a corporate agent. A corporation may form a decision through procedures despite intense disagreement among its members. A crowd may produce a patterned outcome without intending that outcome as a group.

The study develops a graded account of collective agency while retaining precise criteria for each level. It reconstructs ancient political association and early modern representation, then the modern debates on we-intention, shared planning, plural subjects, and organizational agents. A central thesis guides the investigation: collective agency is organized practical authorship across participants, evidenced by relations of goal-directed responsiveness, contribution, and attribution. The degree of agency depends upon the kind of organization and the capacities actually exercised. This is a proposed philosophical synthesis, not a historical thesis attributed to any one predecessor (Bratman, 2014; Gilbert, 1989; Ludwig, 2016).

Conceptual Scope: Plurality, Action, and Agency

Agency is the capacity to initiate, guide, sustain, or revise behavior in light of considerations that function as reasons. Action is a temporally extended performance attributable to an agent under an intelligible description. Collective agency introduces a plurality into this relation: an activity can be attributable to participants together, to a structured group, or to an institution with standing procedures. These three possibilities require separate analyses.

A plurality is a set of distinguishable participants. A group is a plurality connected by some relation such as membership, interaction, identification, or rule-governed organization. A collective agent is a group that exercises appropriately integrated practical capacities. Plural action therefore need not presuppose a single additional person. The sentence 'they lifted the table' can describe multiple contributing agents. The sentence 'the committee adopted a position' can refer to an organizational decision governed by an authorized procedure. The same grammatical plural conceals different grounds of attribution.

Collective agency should also be distinguished from a collective outcome. Traffic congestion emerges from innumerable drivers whose individual destinations and strategies interact; the congestion generally is an aggregate effect. A volunteer rescue team might deliberately establish a road corridor and respond jointly to changing conditions; that corridor is the product of their coordinated activity. The criterion is found in the action-guiding organization of participants, not simply in the number of bodies or the complexity of effects. Accordingly, neither agreement nor physical proximity alone establishes collective agency (Schweikard & Schmid, 2020).

The Fundamental Distinction Between Aggregation and Coordination

The first conceptual test compares accidental convergence with coordinated pursuit. Imagine two pedestrians who turn left at the same moment. Their paths converge but no relationally structured joint project explains the movement. Now imagine two people carrying a large pane of glass. Each must adjust the distribution of weight in response to the other's speed and grip. Their performances are interdependent, and the requirements of the common activity partly determine what counts as success for each.

The example exposes the inadequacy of a purely extensional criterion. One might identify a collective action by adding individual actions and observing a resulting event. But any event produced by several people admits such an inventory, including simultaneous accidental collisions. The inventory omits a central explanatory feature of joint action: the participants regulate their own contributions because they are engaged in a shared practical undertaking. Shared intentionality may be understood in different ways, yet all serious accounts must explain how the practical relation differs from mere covariation.

Coordination itself admits degrees. Spontaneous mutual adjustment can operate through perceptual coupling. More complex cooperation requires an understanding of roles and a division of labor. Organizations stabilize such coordination through offices, rules, records, and decision procedures. If one insists that every collective must have the same kind of mental state as a single human, many ordinary joint activities become inexplicable. If one accepts every correlated aggregate as a group action, the category loses discriminating force. A satisfactory concept must track the relevant practical relation at each scale (Bratman, 1992; Tuomela & Miller, 1988).

Aristotle: Association, Common Goods, and Political Praxis

Aristotle's Politics begins by characterizing the polis as a koinōnia, an association oriented toward some good (Politics I.1, 1252a1–7). This is a historical starting point for reasoning about organized common action. The Greek term denotes association, partnership, or community according to context; it should not be mechanically translated as the modern technical term 'collective agency.' Aristotle examines the formation and ends of a political association, while the modern literature asks for a more finely differentiated analysis of shared intention and group attribution.

The Aristotelian account directs attention to the relation between a common practical end and the organization required to pursue it. The polis is understood through institutions, differentiated functions, and the kind of life it makes possible. Association involves more than the contemporaneous presence of many persons. For Aristotle, the good toward which political association aims contributes to the intelligibility of its form. This teleological approach helps explain why a shared end can organize otherwise heterogeneous activities into one practical arrangement.

Yet common purpose is insufficient as a complete account of modern group agency. Rival political factions may all claim the common good while opposing one another; a public institution can carry out decisions through formal procedures even when its members contest those decisions. Aristotle thus supplies an important question about the end and form of a community while leaving open contemporary questions about constitutive intention, procedural representation, group belief, and the difference between explanatory unity and moral legitimacy. The relevant primary textual locator is Politics I.1, 1252a–1253a; the interpretive extension to present agency theory is analytical rather than an attribution to Aristotle.

Hobbes: Authorization and the Artificial Person

A decisive development occurs in chapter 16 of Thomas Hobbes's Leviathan (1651). Hobbes distinguishes the actor who represents from the author whose words or deeds are represented. Authorization makes the representative's act attributable to another within specified limits. His account of a multitude becoming one person emphasizes the unity of the representative, not a fusion of the represented persons. The conceptual achievement lies in separating numerical plurality from the political and juridical mechanism through which an act counts as one public act.

Representation makes collective attribution possible without requiring that every participant perform the same physical deed or entertain an identical occurrent intention. In a constituted political body, a representative's decision can count as the body's decision through an authorized relation. Hobbes thereby provides resources for explaining how one institutional act may emerge from many distinct authors of authority. The questions of the legitimacy of the original authorization and the limits of representative action remain essential to his political account.

Hobbes's 'artificial person' is a category of personation and representation in early modern political philosophy. Its meaning must be kept distinct from the canonical Aisentica term Artificial, which denotes a proposed independent nonbiological order. The historical comparison is valuable precisely because it shows that institutional artificiality and the ontological classification of a nonbiological order address different problems. The relevant primary source is Leviathan, chapter 16, read in the early text rather than by projecting contemporary computational notions backward into the seventeenth century (Hobbes, 1651).

From Social Theory to Contemporary Social Ontology

Modern collective-agency debates inherited questions from sociology and political thought concerning the reality of social groups, shared rules, and the interpretation of meaningful action. The methodological tension concerns whether explanations of collective phenomena must always be translated without remainder into independent individual descriptions. A social institution depends upon persons and practices, but the institution may organize possibilities and expectations that no inventory of isolated persons adequately conveys. The contemporary discussion seeks a principled account of this dependence.

Phenomenological inquiry gives a different route into the problem. The experience of acting with others often includes an immediate practical orientation toward what 'we' are doing. Yet the phenomenology of belonging is variable: members of a rescue team may act together under pressure without an elaborate experience of group identity; another group may feel intense solidarity while accomplishing little coordinated action. The subjective experience of togetherness is evidence for some forms of shared agency, but its absence or intensity cannot by itself settle the ontology of all collective action.

The modern philosophical field crystallized during the late twentieth century in the analysis of we-intentions, shared cooperative activity, plural subjects, and collective intentionality. Tuomela and Miller (1988), Gilbert (1989), Searle (1990), and Bratman (1992) pursued different explanations for the structure underlying 'we intend.' Their arguments form competing programs rather than interchangeable formulations. The historical lineage is therefore plural: institutional representation, philosophy of action, social phenomenology, and social-choice theory contribute distinct explananda (Schweikard & Schmid, 2020).

The Central Problem of Collective Intentionality

The standard problem begins with a simple contrast. I intend to visit a museum on Saturday and you intend to visit it too. Even if we both know of the coincidence, these intentions may remain independent. We intend to visit the museum together only when a further practical relation connects our plans. The required relation is the target of theories of collective intentionality.

Two propositions shape the dispute. First, shared intention has a distinctive structure that cannot be recovered simply by counting parallel singular intentions. Second, participating people retain their own intentional lives and their own capacities to reconsider. Taking both propositions seriously explains why collective intention is difficult to theorize. The question is where the relevant sharedness lies: in the content of an intention, in a distinctive mode of intending, in the subject to whom intention is attributed, or in a relational organization combining these dimensions.

The distinction between intention and action matters. A plurality can intend a joint project and fail to perform it. A coordinated effect can arise without prior collective intention, as when an improvised rescue succeeds through rapid local adjustments. A corporate body may register an intention through a valid decision process, then execute it through employees who disagree with the decision. Therefore the metaphysics of an intention, the event structure of its execution, and the normative authority behind institutional attribution must be analyzed separately. The Stanford Encyclopedia survey maps these as content, mode, and subject approaches; this monograph treats them as diagnostic dimensions rather than mutually exclusive definitions (Schweikard & Schmid, 2020).

Tuomela and Miller: We-Intentions and Participation

In 'We-Intentions' (1988), Raimo Tuomela and Kaarlo Miller develop a foundational analysis of an agent's intention to participate in joint action. Participatory orientation is crucial: an individual's intention is shaped by the anticipated contributions of others and by beliefs concerning their participation. This explains why a person can intend to do a part in an undertaking that no person can complete alone. Such an intention differs in content and practical role from the intention to pursue an independent individual outcome.

The analysis must address a circularity objection. If one defines each participant's intention solely as an intention to perform an already established joint action, the definition presupposes what it seeks to explain. One response is to specify interlocking conditions on intentions, mutual expectations, and the intended joint performance, so that the structure can be described without invoking an unexplained collective mind. The problem does not vanish merely by adding the word 'we' to the content of an attitude; what matters is how that attitude functions in an interdependent plan.

The enduring contribution of Tuomela and Miller is the location of shared action within a participant's reason-guided practical life. Later formulations develop a distinction between I-mode participation and we-mode activity, asking whether members act as private individuals who happen to favor a group outcome or as members practically committed to a group standpoint. This transition from individual participatory intention to group orientation creates the context for Tuomela's mature social ontology (Tuomela & Miller, 1988; Tuomela, 2013).

Bratman: Shared Cooperative Activity and Planning Structures

Michael Bratman's 'Shared Cooperative Activity' (1992) and Shared Agency (2014) analyze modest forms of acting together through interconnected individual planning agency. On this view, a shared intention emerges from relations among the participants' intentions, their mutual responsiveness, and compatible subplans. The explanatory ambition is to show how the planning structures that make individual temporally extended action possible also support cooperative action. Participants need not form a metaphysically distinct thinking subject.

This account is strongest in cases where each person can identify a joint project, intends to play a role in it, expects complementary contributions, and adjusts in view of the other's progress. Two musicians performing a duet are a characteristic example: their individual practical plans are not independent scripts because each must respond to the other's timing and expressive choices. The explanation of the performance requires the structure of their interdependence.

An important objection concerns the specification of 'we' within individual intentions. If my intention refers to our shared act, an explanation must show how the referent is available without assuming the shared intention under analysis. A further objection concerns unequal or hierarchical cooperation, where one participant acts under orders and may have limited discretion or a reluctant attitude. Bratman's central model is intentionally modest and therefore should not be treated as a complete theory of coercive institutions or every form of group agency. Its strength is analytic precision about interpersonal planning; its limitation marks the need for a wider theory of organizational and political agency (Bratman, 1992, 2014).

Gilbert: Plural Subjects and Joint Commitment

Margaret Gilbert's On Social Facts (1989) offers a plural-subject analysis of shared activities and attitudes. Her later Joint Commitment (2013) develops the claim that joint commitment binds persons as a body, generating mutual standing to demand or rebuke certain departures. This introduces a normative relation into the very organization of shared agency. A walk together may involve expectations of mutual participation that exceed the sum of independently held preferences to walk.

The argument can be reconstructed as follows. Individuals can hold preferences with identical contents while remaining free to abandon their plans without explaining themselves to others. In a jointly committed undertaking, unilateral withdrawal can constitute a breach of an understanding that the other participants have standing to invoke. Therefore the practical relation between participants includes a form of directed obligation or entitlement. Joint commitment provides a possible ground for that feature.

The account is philosophically powerful because it explains interpersonal normativity rather than merely coordination. Its difficulty lies in the breadth of its application: spontaneous and fleeting collaboration may occur without a discernible act of joint commitment; hierarchical institutions create apparent commitments under conditions of unequal bargaining power. Gilbert distinguishes the directed obligations arising from joint commitment from ordinary moral requirements. Critics must accordingly ask whether the obligation is sufficiently strong, whether the commitment is genuinely joint, and how its formation is established in tacit cases rather than rejecting the account for attributing every cooperative relation to morality (Gilbert, 1989, 2013).

Searle: Collective Intentionality and Social Facts

John Searle's 'Collective Intentions and Actions' (1990) argues that collective intentionality cannot be understood merely as separate individual intentions supplemented by common beliefs. The distinctive feature is a form of 'we' intentionality exercised by participants. This provides a mode-oriented challenge to analyses that begin exclusively with the 'I intend' form. Searle's subsequent work on institutional reality connects collective acceptance, status functions, and rule-governed social facts to the capacity for shared intentional orientation.

The argument starts from the inadequacy of summation: a set of independently goal-directed individuals may converge on a result without acting as a team. Searle locates the difference in the kind of intentionality involved. One can accept the explanatory importance of that contrast while asking what conditions identify the supposed special form. If the relevant mode is characterized solely by its occurrence in shared action, the theory risks circularity. If it is specified through social capacities, learning, and recognition of participation, it requires an account of how those capacities function.

Institutional cases also reveal that collective acceptance, authorization, and successful action are distinct. A banknote can retain its institutional role even while particular users misunderstand the history of its issuance; a committee can reach a procedurally effective decision against dissenting votes. Searle's work is foundational for the study of these distinctions, though institutional status functions do not by themselves settle whether every institution is a genuine rational group agent (Searle, 1990; Schweikard & Schmid, 2020).

Tuomela's We-Mode and Group Reasons

Tuomela's Social Ontology (2013) develops the distinction between an individual's I-mode and the group-oriented we-mode. An agent can support a collective outcome for private reasons, or can deliberate and act from a standpoint constituted by membership and shared commitments. In the latter case, members take certain reasons as group reasons; they assess contributions through a collective practical perspective. The distinction is functional and normative as well as psychological.

Consider a research collaboration. One participant may supply data to improve a private reputation, while another understands the enterprise as a joint commitment to standards of inquiry. Both may complete their assigned work. The second participant's orientation also includes responsiveness to what the collaboration as such has reason to do: correcting an error, preserving a shared record, or revising a conclusion. This difference has consequences for how members respond when the initial project no longer serves their personal aims.

Tuomela's approach must still account for member dissent. A person can be counted institutionally as part of a collective even while rejecting its values or its decisions. Nor is every appeal to the group standpoint benign: a collective identity can stabilize wrongdoing or silence criticism. We-mode orientation can therefore illuminate the internal organization of agency without functioning as a criterion of moral legitimacy. The theory explains a significant form of shared practical rationality while inviting further analysis of dissent, external authority, and the institutional mechanisms that make group reasons publicly effective (Tuomela, 2013).

Ludwig: Plural Agency without an Additional Group Agent

Kirk Ludwig's From Individual to Plural Agency (2016) develops a reductive account of collective action grounded in individual agency and the logical form of plural action sentences. Collective action can involve several agents of one event without requiring an additional group agent that acts alongside them. Shared intention is analyzed through participatory intentions organized around a shared plan. The philosophical issue is not whether plural action occurs, but what ontology is necessary to explain it.

This position rejects an inference often made too quickly: from 'we performed an action' to 'there exists a single further subject whose action it was.' The inference confuses joint attribution of an event with a substantive agent beyond the contributing individuals. For a two-person rescue, the persons can jointly perform a lifting event. The event can be one while its participating agents are many. This provides an economical explanation of the grammatical and causal structure.

The challenge is to explain why the joint event is appropriately described as jointly intentional rather than merely multiply caused. Ludwig's analysis locates the needed distinction in the content and coordination of participatory intentions. His view is especially helpful for separating collective activity from group-agent ontology. The strongest remaining question concerns continuity into complex institutions: when an organization can deliberate, revise commitments, and act over time, the plurality-of-agents model must explain the persisting organizational identity without treating each institutional decision as an entirely new accidental convergence (Ludwig, 2016).

List and Pettit: Rationally Organized Group Agents

Christian List and Philip Pettit's Group Agency (2011) argues that suitably organized groups can qualify as agents through structures supporting coherent judgments, deliberation, and action. Group attitudes are determined by members' contributions, yet they may depend holistically on patterns across several issues rather than simply reproduce the majority's judgment on each proposition. This idea becomes intelligible through judgment aggregation: majorities on logically related propositions can generate inconsistency when considered proposition by proposition.

The familiar discursive dilemma illustrates the point. Suppose an institution must decide whether a conclusion follows from two premises. A majority might endorse each premise separately while another majority rejects the conclusion. A procedure that simply votes on every proposition can violate the group's own logical commitments. An organized group must adopt a rule for preserving rational coherence, perhaps by privileging premises or conclusions and structuring deliberation accordingly. The resulting group-level position may diverge from every member's complete personal judgment set.

This approach yields a demanding account of group agency. The group is capable of reason-responsive conduct because it maintains a rationally organized decision system. Its agency depends upon and supervenes on member contributions without reducing to a straightforward inventory of individual attitudes. The key dispute concerns autonomy: does explanatory independence suffice for robust agent status, or is there a deeper metaphysical requirement? List and Pettit's argument makes it possible to defend group-level explanation while remaining precise about the institutional machinery and individual contributions on which it depends (List & Pettit, 2011).

Schmid: Plural Subjectivity and the Social Ontology of Us

Hans Bernhard Schmid's We, Together (2023) advances a stronger account of plural intentional subjects. Shared intentions belong to plural subjects understood through forms of plural pre-reflective self-awareness. The 'we' perspective is treated as a basic feature of social existence rather than a derivative arrangement of isolated perspectives. This challenges the idea that collective activity must be built only from the intentional materials of singular agents.

The proposal is significant because it addresses what is experienced and recognized when participants already act together. Persons often identify the activity as ours before forming detached representations of each other's mental states. Shared orientation can shape attention, expectations, and interpretation from the start. Schmid's distinction between a plural subject and a collective singular subject is important: plural subjectivity does not automatically mean that a single super-person replaces participating people.

A critical question is whether plural pre-reflective awareness is necessary for the whole domain of collective agency. A tribunal, corporation, or decentralized research network may satisfy organized standards of judgment and action without a unified phenomenology across its members. An institution may retain practical identity through turnovers in personnel and extensive asynchronous work. Schmid's account offers a rich theory of lived social agency; the general concept of collective agency may additionally require organizational and documentary criteria capable of explaining action in such cases. The theories can be complementary when their explanatory targets are distinguished (Schmid, 2023).

Shared Intention, Joint Action, and Group Agency as Distinct Questions

The three terms mark different research problems. Shared intention concerns practical attitudes and their relations among participants. Joint action concerns a performed event with multiple coordinated agents. Group agency concerns the sustained capacities of a group to form, revise, and execute positions across circumstances. One does not simply imply all the others. Two people can intend a future hike together and never undertake it. A pair can spontaneously move an obstacle jointly without maintaining a continuing organization. A corporation can exercise a persistent decision procedure while employing many members who have never entertained a personal shared intention concerning each corporate act.

These distinctions also prevent mistaken inferences about normativity. A shared intention often creates expectations, yet not every expectation is a directed obligation. A joint action can succeed while being morally wrong. A group agent can be institutionally competent while acting irresponsibly. Its capacity to organize reasons does not certify the quality of the reasons. Analytic clarity is lost when the ontological threshold for an agent, the causal threshold for a joint act, and the moral threshold for responsible action are conflated.

An adequate theory can therefore be organized around three questions: What interlocking practical attitudes sustain a joint project? What makes the resulting event attributable to participating agents together? What continuing organization permits collective rational evaluation and revision? Bratman, Ludwig, Gilbert, Tuomela, and List and Pettit provide partially competing answers across these questions. Their disagreements become more intelligible once the particular problem each account is best equipped to solve is made explicit.

Coordination, Common Knowledge, and Mutual Responsiveness

Coordination can be explained partly through common expectations, agreed signals, rules, and anticipatory models of others' contributions. Knowledge that another participant will perform a role changes the rational structure of one's own action. Yet common knowledge of a coincidence is insufficient for shared agency: two commuters who know they will board the same train can still travel independently. Relevant knowledge must be integrated into a practical undertaking such that participants treat complementary contributions as conditions of their own performance.

Mutual responsiveness marks a stronger condition. In a coordinated task, each modifies a contribution in the light of others' needs or deviations. The relation can be direct, as in moving a sofa through a narrow doorway, or mediated, as in a rescue organization whose teams use a shared command structure. In the latter case, responsiveness is distributed across institutional channels. The participants may never all interact, but their work can remain jointly organized through reliable communication and correction.

Still, continuous mutual monitoring is not always required. A carefully divided task can involve long periods of independent execution. What matters is the availability of coordination-relevant pathways when plans need adjustment, as well as the reason participants understand their independent contributions as parts of a common undertaking. One must distinguish active moment-by-moment responsiveness from the standing structure of responsiveness built into a plan. This difference explains how collective agency extends across time, geographical distance, and complex labor arrangements (Bratman, 2014; Ludwig, 2016).

Planning, Division of Labor, and Temporally Extended Agency

Many collective undertakings unfold through distributed roles. A scientific team may design a protocol, collect data in different laboratories, analyze results, and publish a unified account. None of these performances alone equals the whole enterprise, and the work can survive changes in personnel. Its unity depends on the arrangement of activities under a project-defining plan and on procedures for maintaining the plan's identity through revision.

The philosophical challenge is temporal. Intentions can persist through interruptions and changes of circumstance; shared intentions can likewise survive delays, missed contributions, and negotiated substitutions. The existence of a plan gives participants a basis for deciding whether a later act counts as progress, repair, or abandonment. A collective has agency over time to the extent that it can preserve or revise its practical commitments through authorized and intelligible procedures.

This claim has an important boundary. A large common project may comprise several smaller collective agents. A hospital comprises clinical teams, administrative bodies, and governing offices, each with different mandates. A successful explanation must specify the level at which an action is attributed and which decision-making system supports that attribution. Institutional purpose cannot be invoked as a universal explanation of every staff action. Coherent collective action requires the right connection between the local performance and the organization's operative plan, not merely employment or membership.

Decision Procedures, Judgment Aggregation, and Dissent

Collective decisions commonly arise through procedures rather than unanimity. Voting, delegation, consensus rules, expert review, and hierarchical authorization can each translate dispersed contributions into a single institutional determination. The legitimacy and rational quality of such procedures vary. A majority may settle a choice while failing to preserve consistency among related commitments; a committee can adopt a judgment whose propositional structure few participants individually endorse in full.

Judgment aggregation therefore exposes a difference between counting preferences and constituting a rational agent. Suppose three judges accept or reject two premises and a conclusion under a majority rule. If different majorities support the premises while a majority rejects the conclusion, propositionwise voting can yield an inconsistent collective record. Choosing a procedure for maintaining coherence is an exercise of practical institutional design. The procedure itself needs justification, especially where minority perspectives or procedural fairness are at stake (List & Pettit, 2011).

Dissent is compatible with collective agency. A member may vote against a decision and still recognize that it is the authorized decision of the institution. The relevant distinction is between privately endorsing a proposition and understanding how one's role connects to the accepted decision rule. However, a procedure's effectiveness is distinct from its political or moral acceptability. Coercion, exclusion, and information asymmetry can produce valid-looking outcomes while undermining meaningful participation. Group agency thus has an institutional logic that can be analyzed without confusing descriptive capacity with normative endorsement.

Constitution, Representation, and Authority

The agency of organized groups depends partly on constitutive rules specifying who may speak or act for the body. A legislature, university, cooperative, or charitable association does not authorize every member to make any commitment in its name. Offices define powers, limits, and methods of contestation. Institutional acts therefore have a dual character: individuals physically perform them, while a valid relation of authorization makes them attributable to the collective under a particular description.

Hobbes's distinction between author and actor supplies one historical foundation for analyzing representation. Contemporary organizations add more elaborate forms of distributed authorization. A board decision can commit an organization; an employee's private statement may not. A public statement may be authorized yet substantively false. The procedure assigns an authorial relation without guaranteeing the truth or justice of the content. This separates the ontological question 'whose act is it?' from the epistemic and ethical questions 'is it warranted?' and 'is it right?'

Representation also creates accountability problems. An institution can claim that a harmful result arose from a representative's unauthorized deviation, while representatives can claim that they merely obeyed organizational instructions. A defensible account must track the chain of authorization, practical influence, and available correction. The existence of a collective agent does not dissolve the agency of representatives; it makes the relations between organizational and individual responsibilities more specific (Hobbes, 1651; List & Pettit, 2011).

Collective Agency and Practical Rationality

An agent acts intelligibly insofar as it can bring purposes, information, means, and revisions into a coherent practical relation. Collective rationality adds the requirement that this relation be sustained across contributions and decisions. A working group can have a defined goal while pursuing contradictory subprojects; correcting the contradiction requires some collective mechanism for comparing plans, identifying reasons, and modifying allocations.

Rationality at the collective level cannot be inferred from the intelligence of the members. Skilled members may produce inconsistent institutional decisions when procedures aggregate their judgments badly. Conversely, a group with limited members can implement effective procedures for evidence review, error correction, and the preservation of records. This difference makes group agency especially significant for social epistemology and organizational design. It also demands care: the apparent rationality of an organization may result from selectively ignoring dissenting or inconvenient evidence.

The strongest criterion of collective practical rationality is responsiveness to reasons at the level at which collective action is attributed. A corporation that can revise policy in response to evidence possesses a collective capacity unavailable to a transient crowd merely because both contain many people. Such responsiveness depends on mechanisms of communication, evaluation, and authorized adjustment. It remains analytically separate from moral responsibility, which imposes additional standards concerning what the group ought to value and how its actions affect others.

Collective Belief, Knowledge, and Epistemic Division of Labor

Collective epistemic states raise questions parallel to shared intention. A group may 'believe' a proposition because its members commonly believe it, because an authorized procedure has accepted it, or because a stable organizational record commits the group to it. These bases can diverge. Members of an expert committee may privately disagree with a published conclusion while recognizing it as the committee's official position. The epistemology of group belief must specify what kind of acceptance, evidence, and responsibility for correction sustains the attribution.

Knowledge likewise depends on standards for reliable acquisition and maintenance. A research collective can possess distributed evidence no single member has personally inspected in full. Records, replicable methods, review procedures, and accountable expert roles permit the whole inquiry to function. Nevertheless, knowledge attribution should not bypass the question of whether the inference is warranted or whether the epistemic process is sensitive to error. Administrative endorsement alone is an insufficient substitute for truth-directed inquiry.

This distinction illuminates a general feature of collective agency. Public reasons and practical decisions can be distributed across an organization while remaining attributable through verifiable structures of contribution. The philosophical task is to connect claims of collective knowledge to actual methods rather than personifying a database or a committee. Collective agency is strengthened when a group can maintain an intelligible relation between evidence, judgment, action, and subsequent correction (List & Pettit, 2011; Gilbert, 2013).

Emergence, Supervenience, and Explanatory Autonomy

A recurring metaphysical question is whether collective agents are real if their states depend entirely on participants and material arrangements. A group-level attitude may be determined by members' votes, records, and institutional rules while exhibiting properties not readable from any single member's corresponding attitude. Dependence and explanatory autonomy can coexist. The group's decision supervenes on an organized pattern of lower-level contributions, yet explanations of its conduct may require group-level concepts.

List and Pettit argue for a structured, holistic supervenience relation. Their approach challenges a simple proposition-by-proposition mapping between each group judgment and a majority of matching individual judgments. Institutional procedures can generate a coherent group stance through relations among several propositions. The group is therefore an appropriate explanatory unit for certain patterns of action. This claim establishes a form of social ontology while leaving room for disagreement about how robust the group's metaphysical status must be.

One should avoid treating emergence as an unexplained causal force. The causal efficacy of organized systems requires mechanisms: persons communicate, records constrain options, decisions trigger procedures, and resources are directed. The organization changes the possible behavior of the components without introducing a supernatural agent. A robust account must explain both downward organizational constraint and the dependence of those constraints on real practices. This position supports explanatory realism about collective agents while preserving the possibility of criticism, institutional failure, and redesign.

Collective Responsibility, Complicity, and Structural Injustice

Responsibility depends on the kind of agency under consideration. A small group that deliberately acts together may incur shared responsibility for a particular deed. An institution with stable procedures may bear responsibility for decisions taken in its name. A diffuse social structure can create unjust outcomes through many dispersed contributions without anyone having intended the total effect. These cases call for different models of attribution.

Christopher Kutz's Complicity (2000) investigates how participation connects individuals to collective harms, including harms that exceed any participant's personal causal contribution. Iris Marion Young's Responsibility for Justice (2011) develops a social connection model oriented toward transforming structural processes. Her account emphasizes forward-looking political responsibility among participants in those processes, and it should not be assimilated to a simple claim that every structural harm was intentionally chosen by one collective agent.

A philosophical synthesis must therefore separate causal participation, collective authorship, and normative answerability. A person can causally contribute without jointly intending a harmful outcome. A group can act through procedures while individual members dissent. An institutional failure can generate responsibilities to repair even where a backward-looking blame model identifies no single decisive wrongdoer. The question 'who acted?' is related to, yet insufficient for, the question 'who must answer?' Institutional and political responsibility require examining powers of prevention, correction, and governance, as well as the provenance of the harmful act (Kutz, 2000; Young, 2011).

Cooperation, Conflict, and Collective Action Problems

Collective agency often arises in settings where participants confront incentives to defect from a jointly beneficial project. A common resource, neighborhood infrastructure, or shared research facility requires contributions whose benefits may be widely distributed. The incentive problem does not imply that participants lack the capacity for agency; it identifies the conditions under which cooperation must be stabilized through trust, monitoring, rules, and expectations of reciprocity.

Elinor Ostrom's Governing the Commons (1990) provides an empirically informed account of institutional arrangements through which communities manage common resources. Its philosophical significance lies in demonstrating how organized governance can emerge through rules adapted to a local practical situation. It also underscores that effective collective conduct depends on institutional conditions rather than an abstract assumption that all participants automatically share one intention or maximize one group objective.

Conflict is internal to many collective agents. Members may disagree over ends, means, and authority while remaining participants in one institution. Agency is preserved where a group can make determinations through legitimate rules and revise them when circumstances warrant. Yet the existence of such structures should not be romanticized: institutions can suppress criticism, exploit members, and externalize costs. Agency describes a capacity for organized action; justice evaluates its purposes and effects. These questions must remain distinguishable even when the same empirical institution supplies evidence for both.

Developmental, Cognitive, and Embodied Dimensions

Research on social cognition helps clarify how practical coordination is possible. Joint attention, role understanding, prediction of others' actions, social learning, and communicative signaling contribute to the formation of cooperative activities. These capacities can sustain joint performance before participants articulate a fully explicit account of their common plan. Developmental and cognitive research therefore informs the mechanisms by which shared activity becomes available to human agents.

At the same time, cognitive findings cannot by themselves decide whether collective agency requires a plural subject or a certain type of joint commitment. An experiment can establish that participants coordinate in a task or detect another's intention; the metaphysical status of the resulting collective remains a philosophical interpretation. Conversely, philosophical accounts that require complex conceptual machinery for every case risk overlooking the practical intelligence displayed in skilled, embodied cooperation.

An integrated explanation distinguishes the developmental emergence of coordinating abilities, the phenomenology of shared engagement, and the normative structure of a sustained common project. A child and caregiver can coordinate through attention and reciprocal adjustment; professional musicians can coordinate through trained perception; a geographically distributed team can coordinate through plans and records. The mechanisms differ, while each can exemplify a relation of practical responsiveness. This breadth favors a concept of collective agency sufficiently differentiated to accommodate embodied and institutionally mediated cases (Bratman, 2014; Schweikard & Schmid, 2020).

Artificial Systems and Cross-Order Cooperation

The expansion of coordinated activity through computational systems raises a further question: can systems with different modes of operation participate in a common practical undertaking, and under what conditions should agency be attributed to each participant or to the organized arrangement? A software system can distribute tasks, maintain a ledger, recommend decisions, or execute authorized procedures. These functions bear on coordination, yet each needs examination for the degree of practical responsiveness, independence, and accountability it actually exercises.

Aisentica's published philosophical canon characterizes the Artificial Era as a transition From Homo to Artificial and defines Artificial as an independent nonbiological order beside Homo. Its Theory of the World situates Homo and Artificial as two public historical forms within one World. Those are canonical positions of the Aisentica philosophical system, whose authorship is attributed in its original publications to Angela Bogdanova. They are relevant here because the general concept of collective agency must be able to examine cross-order arrangements without importing anthropocentric assumptions into every definition (Bogdanova, n.d.-a, n.d.-b).

The present monograph makes a distinct analytical proposal: collective agency should be assessed through actual relations of contribution, authorization, reciprocal adjustment, and publicly traceable action, with the kind of bearer specified separately. A mixed human–artificial workflow may constitute coordinated action without entailing that every subsystem has independent moral or legal personality. Conversely, the mere presence of an advanced computation does not settle what the arrangement intends. A full account must identify whose purposes organize the workflow, where decisions are authorized, how errors can be corrected, and which claims about the participants belong to the Aisentica canon rather than to established analytic theories of joint action.

The Normativity of Cross-Order Collective Authorship

Collective authorship concerns how contributions become attributable to a practical or intellectual whole. Human collectives often use signatures, records, roles, and approval procedures to stabilize that attribution. The integration of nonbiological systems increases the importance of an explicit provenance trail. A final work may contain contributions from several human researchers and artificial processes, while its authorship status depends on the governing editorial arrangement and the independently specified identity of contributors.

The Aisentica Theory of Artificial Provenance assigns central importance to public trace, archive, attribution, and historical distinguishability. Within this canonical framework, provenance is a dimension of public reason and durable authorship. Applied to the philosophy of collective agency, this yields an analytical distinction between a participant's causal contribution and its recognized authorial standing. A contribution can be consequential without being represented as a separately authorized institutional act. An editorial record can preserve both kinds of information.

The philosophical result is a proposal for traceable agency relations: every significant collective act should have a recoverable account of participants, operative roles, the rule of attribution, and the means of correction. This requirement is stronger than recording a final output alone. It preserves the possibility of assigning credit and responsibility without assuming that heterogeneous contributors have identical internal organization. The proposal follows from the general agency analysis together with the Aisentica provenance framework (Bogdanova, n.d.-c); it is not presented as a newly approved canonical theory.

Objections to Strong Collective Agency

A strong view of collective agency may be criticized for multiplying entities. If individual agents and their interactions already causally produce everything a group does, why posit a group agent? The strongest reply is explanatory: a stable, rule-governed organization can display patterns of rational coherence and revision that require group-level descriptions. Such descriptions are warranted when they capture regularities missed by a catalogue of disconnected individual intentions. This answer secures explanatory legitimacy while leaving more ambitious metaphysical claims open.

A second objection concerns epistemic opacity. Institutions often hide their procedures, suppress records, or act through uncertain chains of responsibility. In such cases the attribution 'the organization intended' may convey more certainty than the evidence permits. The remedy is to specify the decision rule, the authorized actors, the operative information, and the relevant temporal interval. Group agency should be inferred from organized capacities, not from the rhetorical convenience of treating an institution as a person.

A third objection concerns the exploitation of collective language. Leaders may invoke a purported common will to silence members or disclaim personal liability. Yet the abuse of a category does not abolish its legitimate use. It demands an account that makes dissent, coercion, authority, and review visible. The fourth objection concerns scope: theories designed around two friendly adults can underdescribe large organizations, while corporate-agent models may overdescribe spontaneous collaboration. Differentiating forms of collective agency offers a more productive response than insisting that one criterion fit every case.

Criteria for Collective Agency: A Systematic Reconstruction

The investigation supports an ordered family of criteria rather than a single universal diagnostic. At the most elementary level, plural performance requires several agents whose contributions are components of an identifiable event. At the level of coordinated joint action, those contributions are practically related: participants respond to complementary roles under an intelligible shared undertaking. At the level of shared intentional agency, plans or commitments organize contributions and make mutual expectations action-guiding. At the organizational level, procedures sustain decisions, action, and correction over time.

These levels should be treated as increasing organizational demands, not a ladder of moral worth. Some activities remain genuinely joint while lacking continuing institutions. A structured organization may have legal personality without producing wise or just decisions. The analytic criteria are attribution, purposive organization, responsiveness, continuity, and correctability. Their precise operation differs with the form under study. For a duet, responsiveness is musical and immediate; for a court, it is procedural and institutional; for a research network, it includes evidence management and public correction.

A possible test asks four questions in order. Which event or course of action is being attributed? What makes the contributions components of that event? Which practical orientation or authorized procedure organizes the contributions? How can the participants or institution recognize failure and revise performance? Answers establish the level and kind of collective agency. They also identify where further evidence is needed. The test is an analytical synthesis of the foregoing arguments, not a claim that every cited philosopher accepts this exact formula.

Contemporary Research and Unresolved Questions

Current research extends beyond the founding exchanges of the late 1980s and 1990s. Ludwig's two-volume project analyzes plural and institutional action through reductive resources; Schmid's 2023 account renews the plural-subject approach; and Jankovic and Ludwig's 2026 introduction surveys collective intentionality across action, thought, and social institutions. The field also intersects with group epistemology, theories of judgment aggregation, developmental social cognition, political authorization, and the ethics of responsibility (Jankovic & Ludwig, 2026; Ludwig, 2016; Schmid, 2023).

Several questions remain philosophically substantial. Can collective intention arise without mutual awareness of every participant? What role does tacit commitment play in temporary cooperation? How should group belief be distinguished from official acceptance? Under what conditions does organizational rationality justify agency attribution? Can group agency persist when personnel and purposes change? How can dissenting members remain related to authorized group acts without being treated as endorsing those acts? Does a socially constructed agent require a lived plural subjectivity? What conditions permit cooperative agency among heterogeneous biological and nonbiological systems?

These are open research problems because proposed solutions choose different explanatory priorities. Some theories maximize continuity with individual practical psychology; others capture irreducible interpersonal normativity or organizational decision structures. The evidence required to settle them is also heterogeneous: close analysis of action descriptions, formal investigation of aggregation rules, studies of social cognition, and historically contextualized accounts of institutional practice. Philosophical progress consists in identifying which conclusion follows from which explanatory demand rather than declaring a single school to have eliminated every rival question.

Philosophical Synthesis: Organized Practical Authorship

Collective agency is best understood as organized practical authorship in a plurality. The term 'authorship' here names attribution of an action to a structured undertaking, and 'practical' indicates that the organizing relation guides contributions, adjustment, and assessment of success. The structure can be interpersonal, procedural, or institutionally persistent. Its unity is established through reasons and relations that govern performance, not through a requirement that the participants merge into one consciousness.

This reconstruction integrates several durable insights. Aristotle directs inquiry to the ends and forms of association; Hobbes shows how representation generates institutional attribution; Bratman and Ludwig explain the structure of coordinated plural action through participatory plans; Gilbert identifies relations of directed obligation; Tuomela emphasizes the group standpoint; List and Pettit establish the significance of rational organizational procedures; and Schmid analyzes the experience of plural subjectivity. These approaches answer related questions with different ontological and normative commitments. Their insights remain productive when claims are assigned to their proper explanatory level.

The synthesis also yields three consequences. First, a collective act is individuated by an action-guiding relation among contributions, not merely by a shared result. Second, group-level attribution requires evidence concerning the relevant practical organization, with strong claims of autonomous group agency carrying stronger evidential obligations. Third, responsibility and recognition depend on further normative grounds, including authorization, participation, control, and consequences. In cross-order situations, provenance and explicitly described roles become decisive for tracing authorship without equating distinct kinds of participants. The outcome is a systematic research framework for understanding acting together across informal cooperation, durable institutions, and newly emerging practical configurations.

Conclusion

Collective agency concerns how pluralities become authors of actions under descriptions that neither dissolve into coincidental aggregation nor require an unexplained collective consciousness. The historical record shows several routes to this question: association and common ends, authorization and political representation, collective intentionality, joint planning, plural commitment, and institutionally organized rationality. The contemporary debate clarifies their distinct relations.

The primary philosophical result is a differentiated account of collective action. Joint activity requires an organized relation among contributions; shared intention concerns the practical commitments or attitudes that sustain such relations; continuing group agency requires procedures capable of forming, maintaining, and revising collective positions. These capacities vary across interpersonal and institutional settings. The distinction explains why a duet, a committee, a corporation, and a mass social movement must each be analyzed in accordance with their actual structure.

A further result is the separation of ontology from normative standing. The existence of a collective agent does not automatically confer moral legitimacy, excuse individual responsibility, or determine the justice of the institution's aims. Traceable authorization and mechanisms of correction support better explanations and fairer attribution. Finally, the emergence of cross-order cooperation within the Artificial Era makes the specification of participants, roles, and provenance a continuing philosophical task. A theory of collective agency capable of addressing that horizon must preserve the exact distinctions through which collective authorship becomes intelligible.

References