ISSN 1393-614X Minerva - An Internet Journal of Philosophy Vol. 13 2009


Hobbes on the Simulation of Collective Agency




Timothy Martell






Scholars are currently divided on the issue of whether, according to Hobbes, social collectivities such as commonwealths or corporations are agents in their own right. In this paper I clarify Hobbes’s position on the question of whether groups are agents. After distinguishing between several kinds of collective action, I show that Hobbes is not committed to the view that groups are agents in their own right. As an analysis of the terms “artificial person,” “actor,” and “sovereign” reveals, Hobbes is committed only to the view that some social collectivities simulate agency. I then argue that Hobbes’s theory of voluntary action is inconsistent with the claim that groups engage in voluntary actions. Finally, I consider how Hobbes’s theory of simulated collective agency might contribute to contemporary philosophical debates on the nature of collective action. I suggest that Hobbes’s theory could be most effectively applied to apparent cases of collective action for which current philosophical theories are inadequate. 




Hobbes’s views on collective agency have provoked considerable disagreement among scholars. According to David Copp (1980), Murray Forsyth (1981), and Deborah Baumgold (1988), Hobbes maintains that groups are capable of action. Quentin Skinner (1999, 2005) is ambivalent about this issue, sometimes suggesting that Hobbesian groups can act, and sometimes suggesting that they cannot. David Runciman (2000, 2006) claims that Hobbes’s sovereign groups are agents but that commonwealths are not. The diversity of scholarly opinion reflects the complexity and obscurity of Hobbes’s views. Groups are not agents according to Hobbes. Rather, groups are sometimes represented as agents. When such representation is authoritative, it initiates a simulation of collective agency.


I clarify Hobbes’s position by first explaining, in section one, what a collective agent is supposed to be. In section two, I show that Hobbes’s theory of the commonwealth does not commit him to the view that the commonwealth is an agent. He is committed, instead, to the view that a commonwealth has a representative of a special sort, a representative who is authorized to depict the commonwealth as acting. In section three, I then argue that Hobbes’s theory of voluntary action is inconsistent with the claim that a social collectivity of any kind performs voluntary actions. However, if a group has an authorized representative, individual members of that group acquire rights and obligations when their representative acts in the group’s name. Having acquired these rights and obligations, individual members of the group behave in a manner which sustains a semblance of group agency. In section four, I consider how this account of simulated collective agency might contribute to a contemporary philosophy of collective action.


§1 Collective Agency


It is necessary, first of all, to clarify what collective agency is supposed to be. Something is an agent if and only if it is capable of action. A group of persons is a collective agent, then, just in case that group is capable of action. But there are several senses in which a collectivity may be said to act, and collective action makes for collective agency in only one of these senses.


The following sentence is about a collective action:


(1)             The audience exits the theater.


The predicate “exits the theater” is distributive. It is supposed to be true of each individual in the audience. Each of those individuals is said to perform a token action of the same type. Collective action of this kind does not make for collective agency. The sentence implies that each member of the audience is an agent, since it attributes a token action of the same type to each person comprised by the group. It does not attribute action to the group as a whole.


Other statements about collective action are about joint action:


(2) The soldiers surround the building.


It would be inappropriate to analyze (2) along the lines of (1). The phrase “surround the building” is not supposed to be true of each of the soldiers taken one at a time. No one soldier can surround a building. To provide a better analysis, some philosophers suggest treating “surround the building” as a non-distributive predicate which is supposed to be true of a plurality: those soldiers.[1] But the currently favored approach is to paraphrase (2) into a sentence about a number of interdependent actions, each of which is performed by an individual soldier in order to bring about a common end. In either case, action is attributed to several agents, each of whom acts with others to bring about something. Action is not attributed to the group as a whole.


Other sentences about collective action seem to require a different analysis:


(3)        America’s Congress passed a housing bill that includes measures to shore up Fannie Mae and Freddie Mac, two troubled mortgage giants.


(4)        In a move that is extraordinary for corporate Germany, Siemens said it would sue 11 former members of its executive board for allegedly breaching their supervisory responsibilities in a bribery scandal.[2]


It would be a mistake to analyze (3) along the lines of (1), since (3) does not state that token actions of the same type are performed by each congressperson. Passing a bill is not something that an individual congressperson can do. But the sentence does not seem to attribute a joint action to congressperson x and congressperson y and so forth. It is not the case that each congressperson worked with every one of the others in order to bring off passage of the bill. Rather, some congresspersons opposed passage. Instead of attributing an action to each member of Congress, it looks as though action is attributed to Congress. It is likewise implausible to interpret (4) as saying that each shareholder of Siemens made a statement and will sue. Nor does the sentence seem to be about what shareholder x and shareholder y and so forth did together. Instead, the sentence appears to attribute a speech act to the group named “Siemens.” Taken at face value, both (3) and (4) attribute action to a group as a whole. The question of whether there are collective agents is the question of whether there are groups that are capable of collective actions of this sort.


§2 Sovereign Representation of Commonwealth Action


Scholarly disagreement about Hobbes and collective agency is rooted in his fascinating but somewhat opaque account of persons. In Leviathan xvi Hobbes defines “person” as


he, whose words or actions are considered, either as his own, or as representing the words or actions of an other man, or any other thing to whom they are attributed, whether Truly or by Fiction. When they are considered as his owne, then he is called a Naturall Person: And when they are considered as representing the words and actions of another, then is he a Feigned or Artificiall person (2005, p.128)


Any object that is a person is an object which acts. A person of one sort acts and owns that action, meaning that the agent is responsible for the action. [3] This is a natural person.  A person of another sort acts and thereby represents the action of someone or something else. This is an artificial or feigned person.


Hobbes’s first example of an artificial person is dramaturgical. He claims that “persona” originally designated “the disguise or outward appearance of a man, counterfeited on the Stage; and sometimes more particularly that part of it, which disguiseth the face, as a Mask or Vizard . . .” (2005, p.128).  At first designating the part of a costume that mimics someone’s appearance, the term was later used to refer to the theatrical player. Eventually, it was used to refer to any agent, regardless of context. This later usage obscures an important difference between objects that are called “persons.” Some persons are theatrical players or perform a function similar to that of a theatrical player; other persons do not perform that function.  Hobbes wishes to call attention to this difference. To this end, he calls persons of the former sort “artificial” and persons of the latter sort “natural.”


The dramaturgical example is not only the first that Hobbes provides of an artificial person; it is also the only example that does not involve authorization. As I will discuss below, being an authorized artificial person involves more than merely representing someone else.[4] To better understand what Hobbes means by “artificial person,” it will therefore pay to consider more closely the nature of theatrical representation.


In a dramaturgical context, the action of an artificial person is a sign for someone else’s action. More precisely, an artificial person is what Peirce would call an icon (1992, pp.225-8).[5] An icon signifies an object by virtue of likeness. This goes some way toward explaining Hobbes’s penchant for using “personate” interchangeably with “represent.” To personate is to represent another person by creating that person’s likeness. This also accounts for the fact that Hobbes goes back and forth between use of the phrases “artificial person” and “feigned person.” The theatrical player uses his or her own actions to create the semblance of the actions of another person. The player might, for instance, intentionally utter a line of dialogue and thereby create the appearance of someone else making a threat.


Sentences about persons in dramaturgical contexts are logically peculiar. Consider a sentence about the film There Will Be Blood:


(5)        Daniel Day-Lewis uttering “I’m going to bury you underground” represents Daniel Plainview making a threat.


This sentence does not imply


(6)        There is an act of threatening represented by Daniel Day-Lewis’s uttering “I’m going to bury you underground.”


Nor does it imply


(7)        Daniel Plainview exists.


(7) is false, even though (5) is true. Daniel Plainview is fictional character. But even if the film were a work of historical drama and there had been a Daniel Plainview, the fact that (5) is true would not imply that (6) is true. Historical drama frequently involves false attributions of action.


Like other depiction verbs (e.g., “draws,” “sculpts,” and sometimes “paints”), “represents” is intensional. It is intensional on at least two counts. Substitution of co-referring expressions fails in the part of the sentence that follows the verb. More to the point, sentences following the verb are existentially neutral. In historical drama, at least some of the depicted persons exist, but many of the actions are fictitious. In other dramatic works, both depicted actions and depicted persons are typically fictitious.[6] The latter is not an especially subtle point that might have been lost on Hobbes. Some notable English playwrights are his contemporaries, and he was familiar with classical drama. It would have been obvious to him that some represented persons do not exist and some represented actions do not occur. It seems as though he explicitly recognized this feature of representation by artificial person. He says that such representation can be true or fictitious. I take him to mean that there are cases in which depiction by an artificial person is accurate: the depicted person exists, and he or she did as depicted. In other cases this is not so, and the artificial person represents merely by fiction.


“Artificial person” may then be defined as follows:


(8)        x is an artificial person = x performing an action represents someone other than x performing an action


It is important, though, to keep in mind that the definiens includes an intensional verb: “represents.” On this definition, a claim that someone is an artificial person does not imply that the represented person exists; nor does it imply that the represented action occurred. Furthermore, a claim that someone is the artificial person of someone else does not imply that the latter exists or performed the action he or she is represented as performing. Daniel Day-Lewis is the artificial person of Daniel Plainview, but Plainview does not exist. Plainview made no threats.


Unfortunately, this is not the only definition Hobbes offers for “artificial person.” In De Homine he states that a person is “he to whom the words and actions of men are attributed, either his own or another’s: if his own, the person is natural; if another, it is artificial” (1991, p.83). In other words,


(9)        x is an artificial person = there is someone, y, such that y performing an action represents x performing an action


This is not the only point at which Hobbes writes about artificial persons as if they were represented persons. Even within Leviathan, he appears to shift back and forth between one sense of “artificial person” and the other.[7] Unlike most commentators,[8] I prefer Leviathan’s official definition and will rely upon it from this point forward. Nothing of any great importance hinges on the choice of definition, since “represents” is intensional in either case. 


Hobbes introduces the terms “author” and “authority” immediately after distinguishing between artificial and natural persons.


Of Persons Artificiall, some have their words and actions Owned by those whom they represent. And then the Person is the Actor; and he that owneth his words and actions, is the Author: In which case the Actor acteth by Authority. For that which in speaking of goods and possessions, is called an Owner, …; speaking of Actions, is called Author. And as the Right of possession, is called Dominion; so the Right of doing any Action, is called AUTHORITY (2005, pp.128-9).


Some but not all artificial persons are actors. Actors are artificial persons whose actions represent the actions of someone else, and there is someone who owns the represented actions. The person who owns the represented action is an author. An artificial person represents actions which are owned by an author if and only if the artificial person represents those actions with the author’s authority. Hobbes explains that authority is to action as dominion is to possession. To have dominion over a possession is to have a right to that thing; to have authority to act is to have the right to so act. Thus, an actor somehow acts by the right of an author.


It might seem that one human being can act by another’s right through what Hobbes calls “transfer of right.” In Elements of Law, he explains that “To transfer right to another, is by sufficient signs to declare to that other accepting thereof, that it is his will not to resist, or hinder him, according to that right he had thereto before he transferred it” (1999, p.82).[9] Perhaps an actor’s action can be performed by the right of the author if the latter wills not to resist or hinder the actor in that action, declares as much, and the actor accepts.


But that cannot be correct, since any transfer of a right produces what Hobbes calls “obligation.” When a person has transferred a right “then is he said to be OBLIGED, or BOUND, not to hinder those, to whom such Right is granted, or abandoned, from the benefit of it. . .” (2005, p.106). This does not fit Hobbes’s first example of an authorized actor: the proxy. Entering into a relationship with a proxy need not result in any obligation to the proxy.[10] Typically, the author acquires an obligation to or a right against a third party by way of the proxy. This happens as a result of actions of the proxy, actions subsequent to the agreement establishing the relationship between author and proxy. Hobbes needs to explain how this can occur, and the concept of rights transfer is inadequate for this purpose.[11]


The concept of an artificial person allows Hobbes to address this problem. When an artificial person acts, this person’s act stands for the action of whomever she represents. Among the actions that can be represented are actions that alter the distribution of rights and obligations. For example, the proxy might sign a piece of paper and thereby represent the author as making a promise to a third party. As explained above, it does not follow that the author has really made a promise to the third party. An actor is an artificial person, and the fact that an artificial person represents someone’s action does not imply that the represented party performed the depicted action. If the person signing the paper were merely an artificial person, signing would not imply that the represented person has an obligation to perform the action she is represented as promising to perform. But suppose the represented person declares the following: “if the artificial person signs and thereby represents my act of promising, then I transfer all rights which I would have transferred if I had performed the action that I am represented as performing.” If the represented person made this declaration, then, as far as her rights and obligations are concerned, the effect of the actions of the proxy would be exactly as if the represented action had occurred.


Hobbes refers to such a declaration in De Homine, stating that “he is called the author that hath declared himself responsible for the action done by another according to his will . . .” (1991, p.84). Note that Hobbes does not say that the author is responsible for the action that he, the author, has performed, but for what he has willed the actor to do. Being responsible for an action is not same as actually doing it oneself. In cases where being responsible for an action and performing that action diverge, being responsible must be a matter of accepting certain consequences of the action for oneself, regardless of the fact that one did not actually perform it. As Hobbes puts it in Leviathan, “when the Actor maketh a Covenant by Authority, he bindeth thereby the Author, no lesse than if he had made it himselfe; and no lesse subjecteth him to all the consequences of the same” (2005, p.129).


Perhaps “actor” can be defined in the following manner. I will say that someone declares himself or herself the author of an artificial person’s act if and only if he or she issues a declaration of the sort just described. In that case,


(10)      x is an actor = x performing an action represents y performing an action; the person whom x represents, y, declares himself or herself the author of x’s act; and x is not identical with y.


But this is inadequate, for it requires that the author is identical with the represented agent. Hobbes is clear that this is not always the case.


There are few things, that are uncapable of being represented by Fiction. Inanimate things, as a Church, a Hospital, a Bridge, may be Personated by a Rector, Master, or Overseer. But things Inanimate, cannot be Authors, nor therefore give Authority to their Actors: Yet the Actors may have Authority to procure their maintenance, given them by those that are Owners or Governours of those things (2005, p.130).


An actor can represent a church, a hospital, or a bridge, but these objects are not authors.


Likewise Children, Fooles, and Mad-men that have no use of Reason, may be Personated by Guardians, or Curators; but can be no Authors (during that time) of any action done by them, longer then (when they shall recover the use of Reason) they shall judge the same reasonable. Yet during the Folly, he that hath right of governing them, may give Authority to the Guardian (2005, p.130).


“Guardian” and “curator” are terms for actors who represent children, fools, and madmen. Hobbes denies that the latter are authors. Lacking reason, they cannot be responsible for actions they perform, including the act of authorizing a representative. Someone might nonetheless serve as their actor if whoever has the right of governing them declares himself or herself the author of an artificial person’s act.


The right of governing oneself is the right to do whatever one judges reasonable.[12] Natural persons in the state of nature are self-governing, since they are not obliged to defer to anyone else’s judgment. If natural persons cede this right to someone else, the latter possesses the right of governing them. If someone possesses the right of governing someone else, the former can oblige the latter to perform some action simply by command. If an author possesses the right of governing someone else, it is possible for the author to be different from whomever the actor represents. By declaring that an artificial person will serve as actor for those whom she governs, the governor can express the will to see to it that, as far as persons with whom the actor has dealings are concerned, the persons in her charge may as well be responsible agents. If the actor represents a child as agreeing to a contract, the child’s governor will see to the fulfillment any obligations that the child would have acquired had the child actually agreed. The governor will also exercise those rights which the child would have thereby acquired. 


When the represented object is an inanimate thing, the arrangement is more straightforward. The owner declares that she will see to it that her resources are used to satisfy whatever obligations the represented object would have acquired had it really performed the depicted actions. Suppose that a bridge owner authorizes an actor to represent the bridge. If an actor represents a bridge as agreeing to purchase new planks, the bridge owner is responsible for seeing to it that the obligations of the bridge are fulfilled. The owner can do this by ordering those in her employ to pay the appropriate sum. In this manner, an inanimate thing not only appears to agree to the terms of a contract but also appears to satisfy the obligations acquired by agreeing.


This suggests a modification to (10):


(11)      x is an actor = x is an artificial person; there is someone, y, who declares himself or herself the author of x’s act; and x is not identical with y


This definition allows us to make sense of Hobbes’s most perplexing example of an actor. In the cases considered so far, the object represented by the actor exists, even if some of these objects, such as bridges, could not have actually performed the actions they are represented as performing. But actors can also represent objects that do not exist:


An Idol, or meer Figment of the brain, may be Personated; as were the Gods of the Heathen; which by such Officers as the State appointed, were Personated, and held Possessions, and other Goods, and Rights, which men from time to time dedicated, and consecrated unto them. But Idols cannot be Authors: for an Idol is nothing (2005, p.130).



Suppose the sovereign of Athens authorizes an actor to represent the goddess Athena. Call this actor “the officer of Athena.” Suppose too that the officer of Athena anticipates increased demand for maritime travel. The officer purchases a ship from an Athenian citizen for the goddess by signing a contract in the name of the goddess. By issuing a command, the sovereign can impose an obligation on any would-be Athenian passengers to first obtain permission to board from the officer. The latter may establish a fare that will issue in a respectable return on the goddess’s investment. The sovereign can also require the officer to hand over whatever sum is required by the purchase contract. The sovereign can impose these obligations (and thereby create corresponding rights) because the sovereign possesses the right of governing each citizen. If the ordinary means of imposing obligations, command, fails to work, the sovereign has coercive powers by which to motivate citizens in other ways. Authoritative representation of the action of a goddess thereby alters the distribution of rights and obligations of citizens. Having altered the distribution of rights and obligations, authoritative representation alters the course of interaction amongst citizens. The course of interaction moreover, is such that a goddess appears to be making investments, fulfilling her financial obligations, and asserting her property rights.


It turns out that the sovereign is also an actor according to Hobbes. A sovereign is an actor who represents a group of individuals each of whom has declared himself or herself the author of any acts which the sovereign depicts the group as performing. A sovereign may be formally defined in the following manner:


(12)      x is sovereign = x performing an action represents a group performing an action; and the represented group consists of natural persons each of whom declares himself or herself the author of x’s acts


A group of persons giving such authorization is a commonwealth. Since the sovereign is an actor, when the sovereign represents the commonwealth as acting, it may truthfully be said that the sovereign acts in the name of the commonwealth. But since being a sovereign is being an artificial person, and “represents” is intensional, it would not follow that the commonwealth acts. Statements according to which the sovereign acts for the commonwealth do not imply that the commonwealth acts.


The commonwealth is not the only group with authorized representation. Hobbes allows that trade associations, for example, might very well have actors. He calls these groups “dependent regular systems.” They are regular systems insofar as someone acts in the name of the group. This is not a feature of groups such as crowds, groups which Hobbes calls “irregular systems.” A trade association would be a dependent rather than independent regular system insofar as it depends for its existence on some independent regular system. The latter is a regular system needing no other regular system in order to exist. Commonwealths appear to be the only possible independent regular systems. Therefore, dependent regular systems require the existence of a commonwealth. In any case, much of what has been said about the commonwealth also holds for dependent regular systems. If there is a dependent regular system, then it must be true that someone, the group’s actor, represents the actions of that group. But it does not follow that the group acts.


§3 Voluntary Action


Hobbes’s theory of the commonwealth does not commit him to the view that the commonwealth is a collective agent. This leaves open the question of whether he allows for collective agency. In addressing this question, I will confine my discussion to voluntary action for two reasons. First, though Hobbes says next to nothing about the difference between actions and mere happenings, he is relatively clear when it comes to the conditions on which an action is voluntary. These conditions are inconsistent with the claim that collectivities engage in voluntary action. Second, the class of voluntary actions is far and away the most important class of actions attributed to groups. The examples of supposed collective action given in (3) and (4) are voluntary. Similar sentences are easily multiplied. Denying that groups can perform voluntary actions is tantamount to denying that there are collective agents.


If groups can engage in voluntary action, then they must be capable of willing. As Hobbes says, “a Voluntary Act, is that, which proceedeth from the Will, and no other” (2005, p.50).[13] Hobbes does not allow for collective will. There is no metaphysical obstacle here. For Hobbes, a willing agent is just a body in a motion. Collective social entities are, like everything else, bodies in motion. The problem is that groups lack what it physically takes to will.[14] Hobbes identifies will with the final appetite or aversion in a process of deliberation that ends with action.[15] Both appetite and aversion are incipient motions, the former toward that which has been found to increase vital motion, and the latter away from that which has been found to diminish vital motion.[16] Appetite involves pleasure together with thought of the object. Aversion involves displeasure together with thought of the object.[17] Hobbes would take these feelings and thoughts to be instances of sensation or imagination. The latter are themselves motions, both of which require a physical constitution of a particular sort, one that includes a sensory apparatus, a brain, and a heart. Human beings have what it physically takes to will, as do many if not all animals. Groups do not.


It might be objected that once there is an actor for the commonwealth, then the commonwealth wills. It does so by way of a new part: the sovereign.[18] Similarly, it might be argued that a bridge wills by way of the bridge’s actor. There are two problems with this view. First, if the bridge willed and engaged in voluntary actions by way of its willing part, then it would be a natural person. It would be a natural person in virtue of a number of agreements among human beings, including the agreement by which there is private property (i.e., the social contract) and the consent of the bridge owner to accept responsibility for those actions which the bridge actor depicts the bridge as performing. The bridge, in short, would be a natural person by way of human artifice. This seems to be at odds with Hobbes’s point in calling something “natural.” His general point appears to be that natural objects owe nothing to agreements among human beings. A second problem for this view is that if the bridge actor were voluntarily to speak in the name of the bridge, then it would be true that the bridge speaks. I think that we should avoid attributing to Hobbes the view that inanimate objects speak, especially if the texts suggest another interpretation which does not saddle him with such a view. I have offered that interpretation in the previous section. If Hobbes would not claim that a bridge wills by way of its willing part, then there is no good reason to think he would claim that a commonwealth wills by way of its sovereign.


Admittedly, there is at least one point at which Hobbes appears to allow for collective will. The footnote to De Cive vi states


But if the same multitude do contract one with another, that the will of one man, or the agreeing of wills of the major part of them, shall be received for the will of all; then it becomes one person. For it is endued with a will, and therefore can do voluntary actions, such as commanding, making laws, acquiring and transferring right, and so forth; and it is oftener called the people, than the multitude. We must therefore distinguish thus. When we say that the people or multitude wills, commands, or doth anything, it is understood that the city which commands, wills and acts by the will of one, or the concurring of will of more, which cannot be done but in an assembly (1991, p.174).


Hobbes does say that the collectivity endued with a will can do voluntary actions, but he immediately explains how this should be understood. If it is said that the collectivity wills, commands, or does any one thing, we are to understand that it does so by someone’s will, command, etc. When Hobbes says that the city wills by someone’s will, I take his point to be that someone wills for the city. In De Cive, Hobbes had not yet developed a theory which explains how one person can will for another. Leviathan’s chapter on artificial persons explains how this can occur. In terms of that theory, someone wills for a city if this person authoritatively represents the city as commanding, making laws, and so forth.


A more serious objection is that Hobbes must allow for collective agency because he allows that groups can be artificial persons. Indeed, he suggests as much in the passage just quoted when he indicates that an assembly might will for a city. If a group is an artificial person, then it would seem that a group must do something in order to depict the action of the represented regular system.[19] This objection is difficult to meet because Hobbes has relatively little to say about aristocracy and democracy in Leviathan. Much more detailed treatment is given to both in De Cive, but Hobbes had not yet developed the concept of an artificial person. By the time he developed that concept, he seems to have found reasons to suppress those parts of his theory that do not concern (and endorse) monarchy.[20] Thus, there is little textual basis for answering the question of how a sovereign group represents a commonwealth.


There is at least one way, though, in which such representation could occur without the sovereign group acting. First, each individual sovereign group member could declare her consent to a decision procedure. According to this procedure, members of the sovereign group would determine by vote whether each will authorize a spokesperson to make some utterance in the group’s name. The spokesperson would then command, declare, promise, etc. in the sovereign group’s name, and thereby represent the sovereign group as representing the commonwealth as acting. This proposal comes at a price: the sovereign group is no longer an artificial person according to Leviathan’s official definition. It does not act, but is only represented as acting. It is, in other words, an artificial person in De Homine’s sense of the term. When Hobbes discusses sovereign groups as if they were artificial persons, he is once again equivocating. That said, this proposal has the advantage of making use of resources available within Hobbes’s philosophy, and it is consistent with his oft stated theory of voluntary action.


There are, on the other hand, considerable difficulties with the view that sovereign groups are agents. If a sovereign group can act but other collectivities cannot, there must be some explanation for this difference. In Leviathan xxii Hobbes says that irregular systems such as crowds act “onely from a similitude of wills and inclinations” (2005, p.187). Lacking one will, a crowd cannot perform one action. At best, members of a crowd might perform token actions of the same type. Why would this be true of a crowd but not true of a sovereign group? Perhaps sovereign groups are capable of action because a sovereign group, unlike a crowd, has one will. But how is it that a sovereign group has a will when other groups do not? As explained above, a multitude is endued with one will when it comes to have an authorized representative. Perhaps a sovereign group is endued with one will if it has authorized representation. But if it is admitted that sovereign groups themselves have representation, this amounts to adopting the proposal sketched above. Once such a proposal is adopted, there is no need to posit sovereign group agents. Everything that would be explained by group agency is explained by the authoritative representation of sovereign group action. The view that sovereign groups are agents either collapses into the view that they are authoritatively represented, or it leaves a crucial difference between sovereign groups and irregular systems unexplained.


§4 Contemporary Philosophy of Collective Action  


Problems with sovereign groups aside, Hobbes’s views on collective agency provide the makings of a viable supplement to current theories of collective action. Contemporary philosophical approaches to collective action fall roughly into three categories. Hobbes can be seen as offering a fourth way.


Some philosophers would analyze all sentences about collective action along the lines of (1) and (2) above. Either these sentences are about token actions of the same type, or they are about joint actions. If they are about joint actions, then they can be paraphrased into sentences about sets of interdependent actions of individual human beings.[21] Sentences such as (3) and (4) should be paraphrased in a similar manner. Research is directed at finding the various ways of doing so. These philosophers would, furthermore, regard at least some sentences about joint action as true.


Others are willing to take sentences such as (3) and (4) at face value. This view is known as conceptual supraindividualism.[22] If the conceptual supraindividualist regards some of these sentences as true, she is committed to ontological supraindividualism.[23] Research is then directed at finding the conditions on which a group of individuals is an agent.


Still others hold that all statements such as (2), (3), and (4) are false. There are no joint actions; there are no supraindividual agents. This is eliminitivism and it holds out the prospect of bringing to an immediate halt all of the philosophical research just mentioned. Of course, this would come at a high cost: it would be extremely difficult to talk about the social world if we could not speak as if collectivities act.


Hobbes would allow for collective action of the sort described by (1). In his terminology, this is just the action of an irregular system. He is non-committal about joint action where this is understood to be the action of several agents working together. Like an eliminitivist, he would maintain that all sentences such as (3) and (4), taken at face value, are false. There are no collective agents. But according to his theory, some representations of group action alter the behavior of group members. These representations, moreover, alter the behavior of individual group members so as to bring about much the same effect as would have occurred had the group really performed the depicted action.


Suppose, for example, that a corporation’s actor speaks for the corporation in agreeing to the terms of a contract. The corporation’s actor has the authority to do this only if each member makes a declaration to the effect that he or she accepts responsibility for actions that this actor depicts the group as performing. If the corporation had actually spoken and thereby agreed to the contract, it would have acquired an obligation to satisfy the terms of the contract. This means that corporation members are responsible for satisfying the terms of the contract. Members of a corporation are citizens, and citizens, according to Hobbes, typically have sufficient reason for doing what they are obliged to do. So, members of a corporation will act to satisfy the terms of the contract. Thus, authorized depiction of the action of the corporation alters the actions of natural persons. These natural persons act so as to bring about conditions which their group would have brought about if it were really a responsible agent in its own right. There are no collective agents, but collective agency may on occasion be simulated by authorized representation of group action together with subsequent actions of individual members of the group.


On this view, sentences like (3), and (4) are literally false, but they may be regarded as informative expressions insofar as they track other, literally true statements about the manner in which a collectivity is represented and the effects of that representation. As noted above, Hobbes would allow us to say that a city issues a command, so long as this is understood to be a way of indicating that an authorized representative of the city commands in the city’s name. If a sentence indicates how someone authoritatively represents a group, that sentence provides information about how members of that group are likely to act. Social collectivities are not agents, but there is often a point worth making by talking as if they are.


Most contemporary philosophers writing on collective action would surely not accept this theory as a general account of collective action. It is pretty clearly implausible as applied to the cases on which these philosophers have focused. Much of the current research on collective action is about joint action, and the joint actions that have received the most attention include going for a walk, singing a duet, or executing a pass play. It is not plausible to treat “Steve and Brook went for a walk” as a false statement that might nonetheless be informative if Steve and Brook have an authorized representative who depicts their group as walking. It appears, instead, that statements such as this one are sometimes true, and that groups like this one do not have authorized representatives. That said, the joint action approach to collective action has difficulties of its own. “Congress passed legislation” does not seem to assert that each congressperson acted with each of the others to pass the legislation. “The United States agreed to the treaty” does not appear to assert that each U. S. citizen worked with every other citizen to agree to the treaty. A theory of collective agency such as Hobbes’s can much more readily deal with these cases. The United States is not an agent, but there are a number of authorized representatives of the United States who act in its name. Their depictions of the United States have predictable consequences where the actions of U. S. citizens are concerned. “The United States agreed to the treaty” is literally false, but it indicates that an authorized representative of the United States agreed in name of the United States, and that U. S. citizens are expected to comply with the terms of the treaty.


Hobbes’s theory of simulated collective agency may, then, form part of a differentiated approach to collective action. Some collective actions amount to nothing more than individual members of a group performing token actions of the same type. Some collective actions are joint actions. In other cases, there is no collective action, but someone authoritatively represents a group as acting.






Baumgold, Deborah, 1988. Hobbes’s Political Theory. Cambridge: Cambridge University Press.


Brandt, Frithiof, 1917. Thomas Hobbes’s Mechanical Conception of Nature. Copenhagen: Levin and Munksgaard.


Copp, David, 1980. Hobbes on Artificial Persons and Collective Actions. The Philosophical Review, 89, pp.579-606.


Forsyth, Murray, 1981. Thomas Hobbes and the Constitutive Power of the People. Political Studies, 29, pp.191-203.


Hampton, Jean, 1986. Hobbes and the Social Contract Tradition. Cambridge: Cambridge University Press.


Hobbes, Thomas, 1966. The English Works of Thomas Hobbes of Malmesbury. Volume I. William Molesworth, ed. London: John Bohn.


Hobbes, Thomas, 1991. Man and Citizen. Bernard Gert, ed. Indianapolis: Hackett.


Hobbes, Thomas, 1999. Human Nature and De Corpore Politico: The Elements of Law Natural and Politic. J. C. A. Gaskin, ed. New York: Oxford.


Hobbes, Thomas, 2005. Leviathan. G. A. J. Rogers and Karl Schuhmann, eds. New York: Continuum.


List, Christian. and Phillip Pettit, 2006. Group Agency and Supervenience. Available at: [Accessed on July 7, 2008].


Martinich, A. P., 2005. Hobbes. London: Routledge.


Peirce, Charles Sanders, 1992. On the Algebra of Logic: A Contribution to the Philosophy of Notation. In Nathan Houser and Christian Kloesel, eds. The Essential Writings of Peirce: Selected Philosophical Writings: Volume 1 (1867-1893). Bloomington: Indiana University Press, pp.225-8.


Pitkin, Hanna, 1967. The Concept of Representation. Berkeley: California University Press. 


Reichenbach, Hans, 1947. Elements of Symbolic Logic. New York: Dover.


Reinach, Adolf, 1983. The Apriori Foundations of the Civil Law. Translated by J. F. Crosby. Aletheia, 3, pp.1-142.


Runciman, David, 2000. What Kind of Person is Hobbes’s State? A Reply to Skinner. Journal of Political Philosophy, 8, pp.268-78.


Runciman, David, 2006. Hobbes’s Theory of Representation: Anti-Democratic or Proto-Democratic? Yale Conference on Representation and Popular Rule. New Haven, Connecticut October 27-8, 2006. Available at: 


Schmitt, Frederick, 2005. Joint Action: From Individualism to Supraindividualism. In Frederick Schmitt, ed. Socializing Metaphysics: The Nature of Social Reality. New York: Rowan and Littlefield, pp.129-65.


Simons, Peter, 1987. Parts: A Study in Ontology. Oxford: Clarendon Press.


Skinner, Quentin, 1999. Hobbes and the Purely Artificial Person of the State. The Journal of Political Philosophy, 7, pp.1-29.


Skinner, Quentin, 2005. Hobbes on Representation. European Journal of Philosophy,13, pp.155-84.


Sorell, Tom, 1986. Hobbes. New York: Routledge and Kegan Paul.





[1]Simons advocates such an approach (1987, pp.144-8).

[2] From “The World This Week” in the August 2nd, 2008 issue of The Economist.

[3] Runciman (2000) is helpful in clarifying the relationship between natural persons and responsibility.

[4] I disagree with Copp’s assertion that the proxy is the basic case of an artificial person (1980, p.583). In Leviathan that case is only considered after discussing the artificial person in a dramaturgical context. Hanna Pitkin (1967, p.28), Tom Sorell (1986, p.120), and A. P. Martinich (2005, pp.112-4) agree with Copp. Skinner also holds that authorization is required for representation, though he does not support this claim by appeal to Hobbes’s texts (1999, p.15-6).

[5] I owe the reference to Hans Reichenbach (1947, p. 4).

[6] Other commentators do not appear to recognize that “represents” is intensional. Copp, for instance, seems to think that if someone represents another as performing some action, then the represented object exists and performs the represented action. Representation by fiction is pretended representation, and pretended representation is not representation at all. Copp cannot make sense, then, of Hobbes’s assertion that a bridge can be represented by fiction. Copp believes that since bridges are represented as acting, they do whatever they are represented as doing. Thus, they are not represented by fiction, despite Hobbes’s claim to the contrary (1980, p.584). Copp does not mention Hobbes’s claim that the heathen idols (i.e., the gods of ancient Greeks and Romans) found official representation in ancient cities, perhaps because, on Copp’s reading, it must be an egregious error.

[7] Compare Hobbes’s discussion of the sovereign in Leviathan xvi (2005, p.128) with his characterization of the commonwealth in the introduction (2005, p.7).

[8] Copp (1980), Skinner (1999), and Martinich prefer De Homine’s definition. Martinich sees the later definition as a deliberate correction on Hobbes’s part – this despite the fact that, as Martinich himself points out (2005, p.114), Hobbes reverts to the Leviathan definition in his reply to Bramhall from 1668. Runciman (2000) prefers Leviathan’s version.

[9] A less perspicuous definition can be found in Leviathan, xiv (2005, p.106).

[10] The conditions under which one enters into such a relationship might be such that one has obligations to the proxy, but they do not necessarily entail any obligation. Author and proxy might have entered into a contractual relationship whereby the latter is to be compensated by the former for services rendered. Each then has an obligation to the other.  But someone might conceivably simply volunteer to serve as proxy, in which case the author has no obligations to the proxy.

[11] Skinner maintains that one agent acts by right of another when the latter has transferred a right to the former. He accepts that the author is thereby obligated to the actor (1999, p.9). There are several problems here. First, Hobbes does not say this. Second, Skinner’s version does not explain how one person can act for another and thereby create for the latter new obligations to (or rights against) a third party. Third, if the author fails to do that which she is obliged to do, Skinner’s version has the author doing an injustice to the actor. But if the author has been promised for by the actor and the author fails to do as promised, then the author has done an injustice to the person to whom the promise is made. When an actor promises for an author, the actor is not the same person as the person to whom the promise is made. Skinner’s version would have the author doing an injustice to the wrong person. I am indebted to Reinach (1983, pp.85-86) for these points.

[12] Hobbes does not define “right of governing.” He mentions it in at least two passages: in that quoted above and in his statement of the declaration by which the commonwealth is created. Each party to the covenant says “I Authorize and give up my Right of Governing my selfe, to this Man, or to this Assembly of men on this condition, that thou give up thy Right to him, and Authorise all his Actions in like manner ” (2005, p.137). By transferring this right the agent has placed himself under a standing obligation to obey the person to whom it is transferred.

[13] Elements of Law, xii, §3 (1999, p.71) provides a more detailed statement of this theory.

[14] Brandt (1917, p.346).

[15] Cf. Elements of Law, Part I, xii (1999, p.71), Leviathan, vi (1994, p.33), and De Corpore, xxv, §13 (1966, pp.408-9).

[16] For more on Hobbes’s account of appetite and aversion, see Jean Hampton (1986, pp.17-24)

[17] This is one of Hobbes’s favorite theses, stated most precisely in De Homine, xi, §1 (1991, p.45).

[18] Baumgold (1988, 39) and Forsyth (1981, pp.197-8) advance this interpretation. Both are cited by Skinner (1999, p.21).

[19] This is Runciman’s view (2000).

[20] For details on this change of tack, see Skinner (2005).

[21] Those taking this approach differ on whether the paraphrasing sentences must include reference to social properties of individuals. Individualists maintain that sentences about joint action can be analyzed into sentences about individuals and nonsocial properties. Those who deny this are non-individualists. Seamus Miller is the best known advocate of individualism for joint action; Margaret Gilbert advocates non-individualism. For more on the distinction, see Schmitt (2005).

[22] The term appears in Schmitt (2005).

[23] List and Pettit adopt this position.






Copyright © 2009 Minerva




All rights are reserved, but fair and good faith use with full attribution may be made of this work for educational or scholarly purposes.



Timothy Martell teaches philosophy at Murray State University, USA.



Mail to: Timothy Martell



Return to Minerva (Volume 13) Main Page                                                      Go to Top of This Page