The Problem of the Unity of the Representative Assembly in Hobbes’s Leviathan

In: Hobbes Studies
View More View Less
  • 1

In Leviathan, Hobbes embraces three seemingly inconsistent claims: (1) the unity of a multitude is secured only by the unity of its representer, (2) assemblies can represent other multitudes, and (3) assemblies are, or are constituted by, multitudes. Together these claims require that a representative assembly, itself, be represented. If that representer is another assembly, it too will need a unifying representer, and so on. To stop a regress, we will need an already unified representer (i.e., one that is not an assembly). But a multitude can only speak or act through its representer, and an assembly is a multitude, so any representing done by the assembly is actually done by this already unified, regress-stopping representer. That is, if (1) and (3) are true, (2) cannot be. I will argue that this inconsistency is only apparent and that we can resolve it without rejecting any of these three claims (and so without imputing error to Hobbes). We do this by appealing to a representer-as-decision-procedure meeting certain criteria. Such a procedural representer breaks the transitivity of representation such that the assembly it represents (and unifies) can properly represent (and unify) some further multitude. I proceed in my defense of the procedural representer view by addressing a series of problems it faces, the solutions to which give us a progressively clearer picture of the criteria this representer must meet.


In Leviathan, Hobbes embraces three seemingly inconsistent claims: (1) the unity of a multitude is secured only by the unity of its representer, (2) assemblies can represent other multitudes, and (3) assemblies are, or are constituted by, multitudes. Together these claims require that a representative assembly, itself, be represented. If that representer is another assembly, it too will need a unifying representer, and so on. To stop a regress, we will need an already unified representer (i.e., one that is not an assembly). But a multitude can only speak or act through its representer, and an assembly is a multitude, so any representing done by the assembly is actually done by this already unified, regress-stopping representer. That is, if (1) and (3) are true, (2) cannot be. I will argue that this inconsistency is only apparent and that we can resolve it without rejecting any of these three claims (and so without imputing error to Hobbes). We do this by appealing to a representer-as-decision-procedure meeting certain criteria. Such a procedural representer breaks the transitivity of representation such that the assembly it represents (and unifies) can properly represent (and unify) some further multitude. I proceed in my defense of the procedural representer view by addressing a series of problems it faces, the solutions to which give us a progressively clearer picture of the criteria this representer must meet.

In Chapter xvi of Leviathan,1 Hobbes introduces a principle for unifying multitudes of people: For a multitude to be unified as a corporate body (call such a unified multitude a civil person, following Hobbes’s usage in De Cive),2 the multitude must be represented by some unified representer.

A multitude of men are made one person, when they are by one man, or one person, represented so that it be done with the consent of every one of that multitude in particular. For it is the unity of the representer that maketh the person one. And it is the representer that beareth the person, and but one person, and unity cannot otherwise be understood in multitude.

The multitude must be unified to reduce their many individual wills and judgments to a single will and judgment standing as that of the collective, as a whole. This is crucial to the establishment of a commonwealth, by which, according to Hobbes, we escape the state of nature – a state of unchecked private judgment resulting in widespread social instability. We do so by appointing a representative (the sovereign) to impose a single will and judgment and investing him with absolute authority, on the condition that he provide protection for his subjects.3 But the point generalizes to civil persons of any kind. Without a unified will or judgment on which to act, a group cannot act as one.4

Hobbes maintains that assemblies can represent multitudes, thereby unifying them into civil persons – e.g., when he says “For the representative must needs be one man or more, and if more, then it is an assembly of all [a democracy] or but a part [an aristocracy].”5 But an assembly is, or is constituted by, a multitude – a plurality of wills and judgments – and so stands in need of unification if it is to represent and unify a further multitude.6 If that unified representer is another assembly, it too will need a unifying representer, and so on. To stop this regress, we will need an already unified representer (i.e., one that is not an assembly).

It might appear that Hobbes gives us a solution to this problem when he tells us, “And if the representative consists of many men, the voice of the greater number must be considered as the voice of them all”.7 However, if mere majority opinion is sufficient to unify an assembly, then it should be sufficient to unify the multitude that the assembly is supposed to represent. Given the fact that Hobbes allows democracies, there can be no principled reason to claim that it is not. But that would mean that every multitude is, in fact, already unified (as a democracy), rendering the requirement that a multitude be represented in order to be unified otiose. This is a result that Hobbes would not accept, as is evident, not only in the requirement that a multitude be unified by its representer, but also in the very nature of the covenant whereby the sovereignty is invested in some person:

And in him [the sovereign] consisteth the essence of the commonwealth, which (to define it) is one person, of whose acts a great multitude, by mutual covenants one with another, have made themselves every one the author, to the end he may use the strength and means of them all, as he shall think expedient, for their peace and common defence.8

The covenant is made laterally between each member of the multitude and each other member, rather than with the sovereign, to avoid the possibility of a breach of covenant, on the part of the sovereign, which might undermine the sovereign’s authority over the members of the commonwealth.9 As Martinich puts it:

The danger, as far as Hobbes is concerned, is that the authority of the sovereign would be mediated by the authority of the commonwealth, and this would open the door for some theorist to argue that the sovereign may fail to represent the commonwealth correctly and that the constituents of the commonwealth, the subjects, could make this judgment.10

Were the commonwealth already unified, any covenant installing a sovereign over the commonwealth would be a covenant between the commonwealth (as a civil person) and the sovereign and would offer no guarantee against this threat. And it is precisely this threat that Hobbes was responding to in Leviathan when he insists on the laying down of one’s right to govern oneself and investing that right in the sovereign by a covenant, not with the sovereign but with each other, only thereby unifying the multitude into a civil person.

Skinner has argued convincingly that we should see this as a response to the parliamentarian writers whose political theories offered justification for the English Civil War that led to the execution Charles i and sent his son into exile. The parliamentarians held that the people, as a unified body prior to the institution of any monarch, invested a monarch with the authority to rule and so retained the right to restrict or revoke that authority. This is, more or less, exactly the result we get by supposing that the majority voice of a multitude automatically determines the voice of the civil person composed of the members of that multitude.11 So we should not saddle Hobbes with that view.12 What goes for the multitude to be represented goes for the multitude that is to do the representing: There is no automatic procedure that guarantees the unity of the representative assembly. To be clear, when I say that there is no automatic procedure that guarantees the unity of the representative assembly, I do not mean to suggest that procedures (such as majority vote) play no role in unifying the assembly. (Indeed – as we shall see – I think they play a central role in doing so.) I mean only that an appeal to such procedures, alone, is insufficient to explain the unity of the assembly in a way that blocks the regress alluded to above.

And the regress carries a further worry. The unification of a multitude just is the unification of the several wills and judgments of the members of that multitude into a single will or judgment, and it is this unified will or judgment that the representer represents. So, were a multitude to be unified by an assembly, which is itself unified by a single representer, the multitude would take the judgment (or will) of the assembly as representing itself. But this judgment (will) just is the judgment (will) of the single representer of the assembly. Furthermore, to represent another is to speak or act on behalf of that other.13 But an assembly can only speak or act through its representer (because it is only through its representer that it has a unified will/judgment on which to act). So that representer must be the one that speaks (acts) for the multitude that the assembly purportedly represents. There is nothing left for the assembly to do by way of representing. When a representative cannot speak or act except through its representative (because it is, itself, a civil person), representation is transitive: If A speaks and acts for B (and B cannot otherwise speak or act) and B speaks and acts for C, then A speaks and acts for C.14 Contra Hobbes, assemblies cannot represent. It looks as though Hobbes is committed to an inconsistent triad: (1) the unity of a multitude is secured only by the unity of its representer, (2) assemblies can represent other multitudes, and (3) assemblies are (constituted by) multitudes.15 Call this the problem of the unity of the representative assembly.

We might try to solve the problem by denying one of the claims that generates it in the hopes that the error attributed to Hobbes in doing so is less onerous than the problem of the unity of the representative assembly. For instance, one might think that a small number of persons constituting an assembly ought not to count as a multitude (claim (3)). But even if we could find some principled means of distinguishing multitudes from non-multitude pluralities (to avoid a sorites sequence), the problem – that there is only unity when there is unity of will and judgment – arises wherever there is more than one judgment or will.16 Denying that the plurality is a multitude is no help. Assemblies of whatever size stand in need of unification. Or perhaps – though they are constituted by multitudes – assemblies only form when their members come together for some purpose. This would make them systems in Hobbes’s terminology. The thought then goes that systems are not multitudes and so are not subject to the principle requiring that multitudes be represented to be unified.17 Systems are either regular or irregular. Regular systems are defined as those that are unified by a representer.18 So, even if the claim that assemblies aren’t multitudes can be sustained on these grounds, it does not help to solve the problem of the unity of the representative assembly: the assembly still must be represented in order to be unified, and it must be unified in order to represent. But what about irregular systems?

Maintaining that representative assemblies can be irregular systems – i.e., “leagues (or sometimes mere concourses of people, without union to any particular design by obligation of one to another, but proceeding only from a similitude of will and inclinations)…”19 – is tantamount to rejecting claim (1): that the unity of the representer is required to secure the unity of the represented. The suggestion is that an irregular system could represent by dividing the authority of the representative among its members, each of whom retains her own judgment and will in determining the exercise of her share of power.20 But Hobbes argues against such a division of power – at least with respect to the sovereignty – on the grounds that it undermines the ability of the sovereign to ensure the security of its subjects that is its very reason for being.21 This reasoning generalizes to non-sovereign authorities, as well: Those powers required to execute the function of the representer form an organic whole and cannot be isolated, one from the other, without undermining the ability to perform that function; nor can the judgment that guides their use be divided without leading to competing factions that interfere with one another’s execution of their duties as representative.

If the office of the representative cannot be borne except by an entity with a unified will or judgment, then irregular systems cannot represent. Given that Hobbes’s purpose, in Leviathan, is to argue for the concentration of power in a single sovereign entity (of unified will and judgment), rejecting the antecedent of the preceding conditional would undermine his whole project. So there is no sense in denying it to solve the (lesser) problem of the unity of the representative assembly. But then, if an assembly represents, it is a regular system – and regular systems require representatives.

This shows us how to mount the objection to the remaining claim in the triad, (2): that assemblies can represent. Given the foregoing, if an assembly cannot both represent and be a regular system, there will be no representative assemblies. And the problem of the unity of the representative assembly gives us precisely the reasoning to argue that an assembly can’t both represent and be a regular system: An assembly can only represent if it is a regular system. And it can be a regular system only if it is represented. But then, by the transitivity of representation, the assembly cannot truly represent. Since rejecting the claim that assemblies can represent leads to no deeper difficulties for Hobbes than that it contradicts his claim that they can, if one of the three claims that generates our problem must be rejected, it is this one.

However, I will argue that we can solve the problem without rejecting any of the claims that generate it while remaining true to the text of Leviathan. We do this by positing a procedural representer that can possess all the necessary features of a representative – e.g., a unified will and judgment and the ability to act – and so serve as the regress-stopping representer of the assembly. This procedural representer is, roughly, a decision procedure that meets certain criteria. I will introduce these criteria sequentially, each as a solution to a particular problem facing the procedural representer view.22 When the assembly is unified by a procedural representer that meets these criteria, there is (I suggest) a break in the transitivity of representation such that the assembly can serve as the proper representer of some further multitude, thereby solving the problem of the unity of the representative assembly without imputing any error to Hobbes. Before turning to these matters, I must make a brief digression to clarify the scope of the problem.

1 The Scope of the Problem

One might hope that, for non-sovereign assemblies, the problem could be blocked by appeal to the sovereign as the regress-stopping representer. This is not correct, though. Consider lawful private bodies, which are constituted for some private purpose (e.g., trade) and operate by permission of the sovereign. The extent of their freedom to act is defined by the sovereign’s law – it is precisely where the law does not dictate what they may and may not do that they have their freedom.23 Since the sovereign has issued no judgment on those matters, we cannot appeal to the sovereign as the unifying voice that stops the regress in these cases.24

Now consider subordinate political bodies. Those that concern us are those that the sovereign has appointed in some representative capacity, either (i) representing the sovereign (e.g., as ambassador) to some other political body or (ii) representing some segment of the subjects of the commonwealth in the capacity of an advisor to the sovereign (as in a parliament). In these cases, one of two things will occur. Either the sovereign stipulates the procedure by which the assembly is to determine its unified will/judgments, or there will be a presumption that the procedure is majority vote.25 But as the assembly members are already subjects of the sovereign, they own and authorize the judgments and actions of the sovereign (according to the sovereign-making covenant),26 both in choosing a procedure and in appointing members of the assembly. Here the sovereign does stop the regress. So, the scope of the problem of the unity of the representative assembly is restricted to sovereign assemblies and private bodies. I now turn to the development and defense of the procedural representer view, by which I hope to solve this problem.

2 Hobbes and Personhood

Hobbes equates representation with the act of personating – bearing another’s person in word and action – and a person with a role. (He cites by way of illustration Cicero’s claim that he bears three persons, his own, his adversary’s, and that of judge, as an example.)27 Persons are further subdivided into two classes – those that act on their own behalf and those that act on behalf of another:

A person is he whose words or actions are considered either as his own, or as representing the words or actions of another man, or of any other thing to whom they are attributed, whether truly or by fiction.

 When they are considered as his own, then he is called a natural person; and when they are considered as representing the words and action of another, then he is a feigned or artificial person.28

The procedural representer view, then, just is the view that there are procedures that can, as artificial persons, represent some (representative) assembly. Lest one think that this is an abuse of the word ‘person’, note that Hobbes distinguishes the person from the physical substance that bears that person. The human (the physical entity) is distinct, even, from the natural person it bears, though this point won’t be relevant to our discussion.29 That the substance bearing the person and the artificial person are distinct is comparatively intuitive: When we consider representative assemblies, we can distinguish them from the collection of individuals that are their members, considered in isolation from any unifying mechanism. But this distinction holds up even where the sovereign is a monarch. The monarch does not, in all his actions, represent the commonwealth. He sometimes acts as a natural person on his own behalf.30

A further distinction amongst non-natural persons must be noted. Not all things represented are capable of acting on their own and yet Hobbes does treat them as persons – by fiction. Hobbes’s examples of such fictional persons are inanimate objects (bridges, churches, hospitals) and humans lacking sufficient reason to count as actors (children and the cognitively impaired).31 The defining characteristic of a fictional person is that it cannot act. If it cannot act it can be neither natural nor artificial person because those roles are defined in terms of representation and representation requires that one be able to act (speaking being a species of action). But recall that civil persons cannot act except through their representer. Civil persons, then, are fictional persons and so cannot be artificial persons (without first being represented themselves).32 This just is the problem of the unity of the representative assembly stated in different terms.

But we can also see a worry for the procedural representer view here. A procedure, to all appearances, lacks will and judgment. If it is a person at all, it is a fictional one and so won’t be able to serve the role I have proposed for it. This line of reasoning rests on the assumption that the only sort of thing that can possess a will and judgment is a rational thing such as a human being. If this is correct, though persons are distinct from the physical bodies that bear them, there is only one sort of entity that can bear them – rational creatures. If the procedural representer view is to work, I must block this result. I intend to do so by rejecting the claim that the only sort of thing that can possess a will or judgment is a rational being (Section 3), by appealing to Hobbes’s theories of deliberation and action.

Rejecting this claim is perfectly consistent with Hobbes’s theory of personhood. The word ‘person’, says Hobbes, is derived from the Latin ‘persona, “signifying any individual thing whatever, acting in whatever way, whether by its own will or by that of another”.33 I see no way to reconcile the claim that a person must be borne by a particular sort of substance (e.g., a human being) with the claim, just quoted, that it is ‘any individual thing whatever’.34 (Given the reference to “acting by its own will or by that of another”, this applies equally to artificial and natural persons.) This supports the possibility of a procedural representer as artificial person.

There is a complication, though. If one interprets ‘acting by another’s will’ to mean that the other’s will just is the will that guides the action and that the purported person has no further will of its own, then lots of unexpected things turn out to be persons (e.g., coffeemakers).35 But Hobbes’s definition of person (quoted above) weighs against this interpretation – it draws a distinction between the words and actions of the representer and those they represent.36 But there is no such distinction to be had between, say, my making coffee (with my coffeemaker) and the coffeemaker’s making coffee (by my will). So coffeemakers and the like are excluded from personhood, which is as it should be. (In the next Section I will explain why the procedural representer is not excluded from personhood on these grounds.)

3 The Problem of Missing Appetites and Opinions

The requirement that a person must be able, on its own, to speak or act would seem to pull against my claim that any substance can bear a person. To assess this worry, we need to get clearer on Hobbes’s theories of speech and action. Here is Hobbes on action: “In deliberation the last appetite or aversion immediately adhering to the action, or to the omission thereof, is that we call the Will, the act (not the faculty) of willing… a voluntary act is that which proceedeth from the will, and no other.”37

This means that our procedural representer must have appetites (and aversions), otherwise it possesses no will and cannot act (and so cannot be an artificial person). A similar issues arises with respect to speech:

Speech [consists] of names or appellations, and their connexion, whereby men register their thoughts, recall them when they are past, and also declare them one to another for mutual utility and conversation, without which there had been amongst men, neither commonwealth, nor society, nor contract, nor peace, no more than amongst lions, bears, and wolves.38

Hobbes calls the train of thought ‘mental discourse’ and the individual thoughts in that train ‘opinions’. “And as the last appetite in deliberation is called the will, so the last opinion in search of the truth of past and future is called the Judgment, or resolute and final sentence of him that discourseth”.39 So in order to speak, one must have opinions – i.e., thoughts composed of things for which we have names. And so just as our procedural representer requires appetites (and aversions) in order to act, it requires opinions to speak; and it must be able to do at least one of these to satisfy Hobbes’s definition of personhood.40 But it isn’t clear how a procedural representer could have either of these required features. Call this the problem of missing appetites and opinions.

As we have seen, Hobbes holds that, “if the representative consists of many men, the voice of the greater number must be considered as the voice of them all” – i.e., the voice of an assembly, by which it expresses its judgment or will, is determined by majority opinion or appetite.41 While we have seen that Hobbes cannot accept this principle as automatically determining the voice of the assembly, he needn’t reject it outright. (He need only qualify it, as we shall see in Section 4.) Furthermore, the principle suggests a solution to the problem of missing appetites and opinions: The members of the assembly contribute their appetites or opinions to the procedure (e.g., majority vote) that determines a single will or judgment that is the will/judgement of the assembly as a whole.42

Not every appetite or opinion of each assembly member will be relevant, of course, so we will require a means of specifying which of these count as proper inputs to the procedure. (Suppose Lum and Abner, proprietors of the Jot ‘em Down Store, have agreed to decide all matters concerning the store by consulting a Ouija board. We don’t want Lum’s opinion that the Ouija board gives better responses when he is seated on its right to be the sort of opinion that impacts the Ouija board’s judgments.) The relevant appetites or opinions must concern the issue before the assembly. For any question brought before the assembly, each member must consider what judgment he thinks the assembly should make or what action she thinks the assembly should take. The judgment (or will) arrived at through this deliberation – i.e., the final opinion (appetite) in his (her) personal deliberation on the question – is input into the procedure as an opinion (appetite) of the assembly. All of his (her) other opinions (appetites) are irrelevant (whether part of the deliberation or not). (If Lum and Abner consult the Ouija board to determine whether they should raise the price of cornmeal, then the relevant inputs are their respective judgments – and only those judgments – on the merits of raising the price of cornmeal.) The procedure operates (‘deliberates’) on these inputs, outputting its judgment (will).

This is not quite enough to get us what we need, though. Consider a vote taken by secret ballot. Those who vote do so according to their own appetites or opinions. If we leave it at that – with a bunch of ballots in the ballot box – we cannot understand the procedure to have spoken. Hobbes tells us that one’s will (and, presumably, one’s judgment) is made known (explicitly) by one’s word or act.43 Until the procedure announces its judgment/will, it remains mute and cannot represent anything. But if the assembly members not only input their wills or judgments but also make them known (anonymously or openly) to the other assembly members – e.g., by having one of their members read out the contents of the ballots in the box – who also know the decision procedure, then we are guaranteed that the outcome of that procedure is known to all the (participating) assembly members. And so the representer will speak. Notice that this places a restraint on the procedure: Holding the inputs fixed, the procedure must yield the same result each time it is run. (There might be inputs other than the judgments/wills of the assembly members – e.g., in the case of a tied vote, the procedure might specify a coin flip as a tie breaker. The outcome of the coin flip would also count as an input.)44

This leaves us with a procedural representer that takes input from human wills or judgments (thereby solving the problem of missing appetites and opinions), communicated to the other members of the assembly (thus allowing the representer to speak). But there are other worries lurking in the vicinity; namely, how to determine who is authorized to carry out the procedure and who, exactly, it speaks for.

4 The Problem of Which Multitude

Recall, once more, Hobbes’s claim that the majority opinion determines the voice of the collective. His rationale is that the majority, being superior in numbers, will cancel out the number of dissenting opinions while leaving at least one voice in favor of its opinion as a remainder.45 As we have already seen, this cannot be the whole story. But there are further qualifications that must be added to majority voice (or any other decision procedure) if it is to serve as a procedural representer. First and foremost, each member of the assembly must authorize the procedure as the procedure for determining the voice of the assembly. This follows from a simple extension, to the procedural representer, of Hobbes’s comments on the representative as the unifier of the multitude – e.g., that a multitude is unified into a single (civil) person when it is represented by a single person “with the consent of every one of that multitude in particular”46 – and quite explicitly from comments he makes concerning the majority’s role in determining the voice of the multitude (at the initial institution of a commonwealth) in De Cive: “…each member of a crowd must agree with the others that on any issue anyone brings forward in the group the wish of the majority shall be taken as the will of all…”47 To illustrate the need for such a qualification, consider the following case.

A group of twelve friends set out on a walk in the country, intending to get some honey for a tea party they will be hosting. Seven of the twelve express a preference for clover honey, whereas five express a preference for mesquite. We are quite content to say that the group speaks in favor of purchasing clover honey. In short order they arrive at Mr. Billingsley’s bee keeping establishment. Mr. Billingsley, being overstocked on mesquite honey, prefers that they purchase mesquite. If majority opinion is sufficient to determine a unified voice and having a unified voice is sufficient to unify a group into a civil person, then there will also be a civil persons comprised of the original twelve plus Billingsley (and this group also prefers clover honey).48 We are not so content with this result. There is some intuitive sense to the twelve constituting a civil person that prefers clover honey, but none to saying that there is any civil person constituted by the twelve plus Mr. Billingsley. This is due to an implicit assumption that the group of twelve decided, together, to purchase whichever sort of honey the majority of them prefers, in which case each member of the group of twelve has authorized the decision procedure as delivering the judgment of the group as a whole. Billingsley, by contrast, has not authorized any such procedure, mutually, with the twelve friends.

So, Hobbes’s rationale for the majority opinion determining the voice of an assembly (or any other group) must be understood as operating against the background of a requisite authorization of that procedure. (The point generalizes to other procedures as well.) If the fact of a majority opinion among any collection of people were sufficient to give them this unity, then unity would be much too cheap (and would lead to the problems we previously noted for Hobbes’s project): Any group of, say, seven individuals sharing an opinion is sufficient to unify every group of thirteen or fewer people (and any six of those seven will be sufficient to unify a distinct group of eleven or fewer people, and so on). This is clearly not what Hobbes intended. Call this problem the problem of which multitude. The problem is solved, in part by this appeal to authorization.

But in the case of a procedural representer, there is another aspect to this problem that must be addressed: identifying who is authorized to give the inputs (of wills and judgments) to the procedure. For the assembly will not only authorize the procedure as delivering their voice, they must also authorize those that are to employ the procedure.49 This is a simple matter for a majority opinion procedure or any other that takes the authorizing persons as those that provide the required appetites or opinions. They need merely recognize each other (and themselves) as authorized to provide these inputs. Where this is not the case, we need an authorization of the procedure and either those who are to provide these inputs or the means of selecting them (see Section 5).50

Incorporating this into our account, the procedural representer must (at least) take input from human wills and judgments (thereby solving the problem of missing appetites and opinions), that are publicized to the other members of the assembly (giving voice to the representer), where the members of the group represented by the procedure have each endorsed the procedure and authorized (each other as) those who are to give inputs to the procedure (thus solving the problem of which multitude).

5 The Problem of Action

We have seen how the procedure can be made to speak to the members of the assembly, but we have not seen how it can act (nor how it can speak to the multitude it purportedly represents).51 Without this, though we will have an artificial person unifying the assembly, we won’t have one that can represent a further multitude by acting (speaking) on its behalf. So we need provisions, not just for authorizing the voice of the assembly, but also for acting upon its will or expressing its judgment to the multitude to be represented.

The answer to this problem is found in an analogy with the sovereign-making covenant that institutes the commonwealth. Hobbes defines the commonwealth as: “one person, of whose acts a great multitude, by mutual covenants one with another, have made themselves every one the author, to the end he may use the strength and means of them all, as he shall think expedient, for their peace and common defence”. He goes on to define the sovereign as “he that carrieth this person” – i.e., the one that represents the commonwealth.52 That is, the people incorporate themselves as and lend their strength and means to their collective (civil) person, to be represented by the sovereign. The sovereign’s judgments are the will of the commonwealth and are carried out by the subjects of the commonwealth, using their pledged strength and means.

By analogous reasoning, if the assembly is going to be able to act or pronounce judgment, then it must not simply authorize a decision procedure to which it inputs its appetites and opinions to produce a single will or judgment. It must also undertake an obligation to treat the procedure’s output as the will or judgment of the assembly (as a unified civil person) and to act or speak (to the body they represent) accordingly, lending to that procedure-as-artificial-person their power to act (in the relevant domain), just as the multitude the assembly represents lends their strength to the assembly. There is an important disanalogy with the sovereign-making covenant, though. In this case there need be no giving up of the right of self-governance, in total, to the procedure. (The significance of this point will become evident in Section 6.)

Here a complication arises. We can understand the relation of the assembly members to the procedural representer in two ways. On the first the assembly members are constituents of the procedural representer. On the second understanding, the representer has offices as constituents, but the assembly members that fill those offices are not, themselves, a part of the representer. Call the first view the officer view and the second the office view. I think there are compelling reasons to prefer the office view.

First of all, the text of Leviathan offers support for the office view: “In an aristocracy, when any of the assembly dieth, the election of another into his room belongeth to the assembly, as the sovereign, to whom belongeth the choosing of all counsellors and officers”.53 Here we have reference to the ‘room’ (office) occupied by an ‘officer’, which room is vacated by the officer’s death. This clearly supports the office rather than the officer interpretation. (On the officer interpretation, we would expect Hobbes to say that, when the officer dies either the assembly continues without him or someone else is elected to replace him – not that someone else is elected into his place.)54

There is also an epistemic worry about the officer view – namely, that from a given moment in time, there are any number of possible sequences of succession for each officership. Since the representer must be unified in order to secure the unity of the multitude represented, only one such succession can be the representative. But then how are the members of the multitude to know, after a change in the constitution of the assembly, whether or not the resulting assembly is still the assembly that was initially authorized as their representative? Without either positing an implausible foreknowledge on the part of the members of the multitude or attributing a de facto theory to Hobbes (which is controversial at best),55 they cannot. (Notice that this problem is not unique to assemblies.) So we have another good reason to reject the officer view.

This leaves us with a procedural representer that is authorized by each member of the assembly it represents and that includes offices, the holders of which (a) are recognized as such by the assembly members, (b) contribute their wills or judgments to the decision procedure (to solve the problem of missing appetites and opinions), (c) know the procedure’s operations and the wills or judgments input by the other officers (to allow the procedure to speak), and (d) have undertaken an obligation to act (or speak) as directed by the assembly, in accordance with the will (or judgment) determined by the decision procedure (thereby solving the problem of action).

6 Solving the Problem of the Unity of the Representative Assembly

The foregoing gives us an account of how the assembly gets unified by way of the procedural representer. This account affords two ways of unifying a further multitude by way of that assembly. In the first, the civil person of the assembly is authorized as the representer of the multitude. In the second, the same mechanism (the procedural representer) that unifies the assembly unifies the multitude to be represented by that assembly. Which of these approaches is taken will have consequences for the constraints the decision procedure places on the assembly. In the former case, since authority is vested in the assembly, the means of determining the assembly’s will (judgment) have no bearing on the obligation the multitude undertakes in authorizing the assembly. And so, even if the assembly alters its decision procedure, the multitude is still bound by its initial authorization of the assembly. In the latter case, however, if the procedure has not been explicitly authorized to make alterations to itself, then the assembly cannot, by means of that procedure, make any changes to the procedure without stripping the procedure of the authorization of the multitude. Or so it might seem.56

This appearance is deceiving, though. The procedure has been authorized by the multitude, and so its determinations, even those concerning alterations to itself, are authorized by the multitude, whether explicit authorization to make such changes has been given or not. However, there remains a crucial – albeit subtle – difference between the two cases; namely, that when the procedure has been authorized by the multitude, any changes to the procedure must be made by way of the procedure. When the assembly has been authorized by the multitude, the assembly members could choose – not by means of the procedure but by unanimous opinion – to replace the existing procedure with another.57 This is allowable because, though there has been an analogy drawn with the sovereign-making covenant (to solve the problem of action), the procedural representer is not, in the present case, the sovereign of the assembly members and so they have not given up their right to govern themselves to that procedure.58 Unanimous opinion, then, is sufficient for the assembly members to (i) transfer back – by means of the procedure – the strength and means they have committed to the procedural representer (to solve the problem of action) and (ii) to release one another of their mutual covenant to abide by the will and judgment of the procedural representer, thereby freeing them to establish a new procedural representer.59 The reasoning for (ii) is straightforward. The reasoning for (i) is that the wills/judgments of the assembly determine (as inputs to the procedure) the will/judgment of the procedural representer. So, where the assembly members are unanimously of the opinion that (i) should obtain, the procedural representer will also be of that opinion, and the procedural representer is the one with the power to cancel the obligation the assembly members have undertaken to lend their strength and means in support of the procedure’s determinations.60

We can now add the last link in the chain of reasoning that blocks the problem of the unity of the representative assembly: If the assembly is directly authorized, then we have good reason to think that the procedural representer is not the true representer of the multitude – despite appearances to the contrary – because it can be changed by the assembly without appeal to the initial procedure (and so without the authorization of the multitude) and yet without an impact on the assembly’s status as representative of that multitude.61 So the problem of the unity of the representative assembly is solved; the procedural representer breaks the transitivity of representation. The assembly can represent without its procedural representer (by which it is unified such that it can represent) being the true representer of the represented.62

Where the procedure is authorized by the multitude directly (which, anyhow, seems unlikely), the problem of the unity of the representative assembly does not arise because it is not a proper instance of representation by assembly. It is an instance of representation by a procedural representer. It’s just that it looks, for all the world, like an instance of representation by assembly because the true representer is only able to function as a representer with the assistance of the assembly (i.e., by taking their wills and judgments as inputs, etc.).

7 Policing the Procedure

One additional worry that is likely to arise concerning the procedural representer is that there are inadequate checks on whether or not the procedure has been properly implemented. What we should say about this depends on whether the assembly or the procedural representer has been authorized as sovereign.63 If the assembly has been authorized (and hence is free to alter the procedure as it sees fit without further authorization from the multitude), then the answer is simple: The assembly will determine whether or not the procedure has been properly implemented, and they will do so via the procedure. Should that determination be called into question by an assembly member, then a further appeal to the procedure to determine the judgement of the assembly will take place, and so on. But the subjects of a commonwealth are not, according to Hobbes, justified in objecting to the pronounced judgments of the assembly even if they are the result of an improper application of the procedure (and so not properly the judgment of the sovereign assembly at all).64 (Since it is the assembly, not the procedure, that is sovereign here, there is no problem about assembly members questioning the procedure.)

In the case where the procedure is authorized as sovereign by the multitude, the judgments and actions of the procedure are only binding so long as the procedure in use is, at that time, the authorized procedure. Where the implementation of the procedure is not fully transparent – especially if its rulings are seen by the subjects of the commonwealth to be unfair – the appearance of injustice is likely to take hold. If the apparent injustice is persistent or pernicious enough, rebellion is likely (though not justified).65 It will be advisable, then, to ensure that the procedure’s implementation is as transparent to the people of the commonwealth as is possible. The practical difficulty of ensuring this transparency to the satisfaction of all citizens of the commonwealth (witness the prevalence of conspiracy theories in our own society) gives us reason – from the Hobbesian perspective – to prefer authorization of the civil person of the assembly over than authorization of the procedure that unifies the assembly.

There is a related worry: Hobbes has said “that it is men and arms, not words and promises, that make the force of the laws”.66 But this worry only gains traction if one supposes that the procedural representer is nothing more than words and promises, and the whole point of the discussion to this point has been to show how a procedural representer could act and speak and, therefore, could count as an artificial person. In particular, we have seen (in Section 5) how the procedural representer could come to act through the assembly members and the subjects represented by that assembly. It is through this mechanism that the laws passed by the assembly (as represented by the procedure) can be enforced. (A monarch, too, must have her army or police force to enforce her laws).

At any rate, Hobbes’s comment is aimed at the view (attributed to Aristotle) that the sovereign should be subject to the laws. Hobbes’s idea is that, if there were no source of judgment over and above the laws – particularly, one interpreting the laws – then the laws could not be enforced. And the source of this judgment, the sovereign, cannot be subject to those laws for there is no one with the power to interpret and enforce them against the sovereign.

But this is not a problem for the procedural representer. The procedure is either itself the sovereign or it is the representative of the sovereign assembly but not the true representative (sovereign) of the commonwealth. In the former case, the worry is that the procedural representer is subject to itself. But it is not at all clear what it would mean for this to be the case. How, for instance, could the process of majority vote be subject to majority vote? The decision to proceed by majority vote could be made by majority vote, but these are not the same thing. So the worry doesn’t get off the ground. In the latter case, the assembly is sovereign and is, therefore, the source of the force of the commonwealth’s laws. There is nothing in my view that requires the assembly to be under the jurisdiction of those laws. If one thinks that the procedural representer, itself, constitutes a law over the assembly, we need only note that – even if it is properly thought of as a law – the procedure is not a law of the commonwealth and so doesn’t bear on the assembly’s ability to give force to the laws of the commonwealth. And there is reason to think that the procedure isn’t properly thought of as a law over the assembly because (as we previously saw) the assembly can replace the procedure with another. So we have diffused the worry for the latter view, as well.

8 Conclusion

The problem of the unity of the representative assembly arises from three apparently inconsistent claims endorsed by Hobbes: (1) the unity of a multitude depends upon the unity of the one that represents it, (2) assemblies can represent (and thereby unify) multitudes, and (3) an assembly is (constituted by) a multitude. These claims seem to be inconsistent because, in order to represent a multitude the assembly must, itself, be unified. But to be unified requires that the assembly be represented, in which case the representer of the assembly would seem to be the true representer of the multitude that the assembly purportedly represents. So the assembly cannot represent the multitude after all.

I hope to have shown that Hobbes’s political theory has the means to resolve the apparent inconsistency without rejecting any of these claims. In particular, I have argued that the artificial person representing the assembly (and thereby unifying it such that it can represent some other multitude) can consist in a decision procedure unanimously authorized by the assembly members in a covenant amongst themselves in which they also undertake the obligation to lend their resources to the realization of the will and expression of the judgments of the collective assembly as determined by that decision procedure. Such procedural representers break the transitivity of representation required to generate the problem of the unity of the representative assembly – the procedural representer allows the assembly to represent a further multitude without the procedural representer being the true representative of that multitude.

Thanks to S.A. Lloyd, Paul Garofalo, and two anonymous referees for this journal for their insightful comments on earlier drafts of this paper.


T. Hobbes, Leviathan: with selected variants from the edition of 1688, ed. with introduction and notes by E. Curley. (Indianapolis in/Cambridge: Hackett, 1994).


Notice that the notion of a civil person is distinct from either that of the natural person or the artificial person as defined in Leviathan, 81. The latter two notions involve representation and ownership of actions whereas the former concerns the incorporation of many into one.


A major purpose of Leviathan is to give us an account of the origins of the state that justifies imbuing absolute authority in the sovereign and so insulating the state from the threat of civil war (which Hobbes likens to the state of nature, Leviathan, 63).


Hobbes discusses this explicitly with respect to the formation of the commonwealth, as represented by the sovereign in both Leviathan, 88 and T. Hobbes, On the Citizen, ed. R. Tuck and M. Silverthorne (Cambridge: Cambridge University Press 1998), 73. The passage from On the Citizen (De Cive) goes on to discuss the civil persons other than the commonwealth.


Leviathan, 94.


This commitment is left implicit in Leviathan. I will argue, below (in the discussion of assemblies as systems), that Hobbes is genuinely committed to this claim.


Leviathan, 82.


Leviathan, 88.


Leviathan, 89.


A.P. Martinich, “Authorization and Representation in Hobbes’s Leviathan, in A.P. Martinich and K. Hoekstra (eds.), Oxford Handbook of Hobbes (Oxford: Oxford University Press, 2016), 335.


Q. Skinner, “Hobbes on Representation,” European Journal of Philosophy 13:2 (2005), 155–184. See also, M.B. Vieira, The Elements of Representation in Hobbes (Boston: Brill, 2009), especially 180–198.


I will discuss further worries with this view in Sections 3 and 4.


Leviathan, 81.


This transitivity, of course, is restricted to some domain in which A is authorized to represent B. And it will not necessarily go through where B is able to speak or act independently. A might speak and act for B in some domain, D, while B is free to act and speak in domain D’. Now C might appoint B to speak and act for it with respect to D’. Clearly, A will not represent C with respect to either D or D’.


So far as I am aware, the closest thing to a discussion of this problem in the secondary literature is the following passing mention from A.P. Martinich in his discussion of Hobbes’s treatment of the trinity: Another problem concerns the ability of the apostles to represent anyone as a group. Since the unity of the thing represented depends upon the unity of the person representing them, the apostles would need to be a unity. But they can only be unified by a sovereign-making covenant, and they clearly are not so unified according to Hobbes’s views. (The Two Gods of Leviathan: Thomas Hobbes on Religion and Politics (Cambridge: Cambridge University Press, 1992), 207.)

Martinich does not generalize the problem to any representative assembly whatsoever, nor does he offer a solution to the problem. However, his suggestion that the assembly (in his case, the apostles) be unified by a sovereign-making covenant is, we shall see, on the right track.


Anyone who has ever attempted to plan a dinner date knows that such unity of judgment and will – even in a group of two – is not easy to come by.


There is a mistake in this reasoning. The mere fact of coming together for a purpose is not sufficient for unity, given the principle regarding the unification of multitudes. This is just what falls out of the discussion, to follow, regarding regular and irregular systems.


Leviathan, 115.


Leviathan, 122.


Jean Hampton employs this reasoning in her argument that assemblies cannot bear the sovereignty. The claim that irregular systems cannot bear sovereign power, together with the assumption that sovereign assemblies must be irregular systems, yields her conclusion. However, she offers no argument for her assumption (which is left implicit in her discussion). The present account will offer reason to reject it and, therefore, to reject her argument. (J. Hampton, Hobbes and the Social Contract Tradition (Cambridge: ­Cambridge University Press), 105–106.).


Leviathan, 86, 93, 171.


Among these criteria will be those that, when met, allow us to avoid the problem raised, above, for the suggestion that mere majority opinion is sufficient to unify the assembly. A majority opinion procedure that meets these criteria will be a perfectly good procedural representer.


Leviathan, 150.


Despite the fact that it is by the sovereign’s authority that the private body is allowed to exist. I will have more to say about these matters in Sections 4 and 5.


It is possible that the sovereign would explicitly leave the choice of procedure for determining the voice of the assembly up to the assembly members. But it is implausible that this would include their choosing some human to act as their representative. Had the sovereign wanted that person as a representative, surely she would have chosen that person rather than the assembly. If choice of procedure has been left to the assembly, then they are in roughly the position of a lawful private body.


Leviathan, 88.


Leviathan, 81; Leviathan Appendix, 328.


Leviathan, 81.


It might seem that, if we reject the identity of the acting substance and the natural person, then we risk losing the distinction between natural and artificial persons. But denying identity does not preclude some other privileged relationship between substance and natural person. What this relationship could be – and whether it or identity is the proper natural person/substance relation – is a problem that exceeds the scope of this paper. Our concern is with artificial and civil persons. Also, see the discussion of persons by fiction, below.


Leviathan, 95, 122–123.


Leviathan, 81.


See D. Runciman, Pluralism and the Personality of the State (Cambridge: Cambridge University Press, 1997), Chapter 2 for a discussion of the civil person – in the form of the commonwealth – as a fictional person. We should also note that Hobbes’s definitions of artificial person in other texts – namely De Homine and the Latin Leviathan – do not accord with the definition offered in the English Leviathan. The point of difference is whether artificial personhood applies to the representer, the represented or both. See Martinich, “Authoriziation and Representation in Hobbes’s Leviathan”, 324–332 for an illuminating discussion. The English Leviathan is quite clear that it should apply to the representer, the Latin text is ambiguous between the two interpretations. Since I am most interested in the English Leviathan, I will accept that an artificial person is a representer and remain neutral on whether we should also count those represented as artificial persons (whether fictional or otherwise). The problem raised here for representative assemblies – that civil persons can only act through their representers – cannot be swept away by a change in a stipulative definition.


Leviathan Appendix, 341.


Certainly Hobbes wanted to reject any suggestion that the representative somehow resemble the represented, a principle that had been put to use by the parliamentarians. See Q. Skinner, “Hobbes on Persons, Authors and Representatives,” in P. Springborg (ed.), The Cambridge Companion to Hobbes’s Leviathan (Cambridge: Cambridge University Press, 2007) and Vieira, The Elements of Representation in Hobbes.


Hobbes suggests – but does not explicitly state – that ‘persona’ is a synonym for ‘person’.


Individuating actions in terms of actors, i.e., whose will provided the immediate impetus to the action.


Leviathan, 28. This passage suggests a distinction between actions and voluntary actions, voluntary actions being those actions that are intentional. Since we are concerned with artificial persons, we can set involuntary acts aside: Whenever one acts on behalf of another one must deliberate as to the interests of the represented party, and so all acts as a representative will be intentional acts. (One cannot, for instance, flinch on behalf of another. Whether or not one can flinch – or perform any other involuntary act – on one’s own behalf is a matter we don’t need to resolve for the present discussion.) For this reason, wherever I speak of actions, I should understood as speaking specifically of voluntary actions (which, after all, just is what we mean by ‘action’ in modern usage).


Leviathan, 12.


Leviathan, 30.


Leviathan, 81.


Leviathan, 82.


This differs from the coffeemaker example (of the preceding section) in the following crucial way: In that case, my will initiated the act of the coffeemaker (if it makes sense to attribute the act to the coffeemaker at all). Were the coffeemaker to operate on the lines of the present proposal, I would press a button to signal my desire for coffee. The coffeemaker then, by its own will, would either satisfy or upset my desire. In this case, there are clearly two acts: my signaling of my desire to the coffee maker, and the coffeemaker’s making the coffee (or not).


Leviathan, 141.


This restriction has the further benefit that it prevents an assembly member from gaming the system by repeatedly bringing some issue before the assembly until he gets the result he prefers.


Hobbes is not entirely committed to the view that the voice of the multitude is determined by majority opinion. First, he allows that split decisions might yet determine some judgment or will. He goes on to consider an assembly where each member has veto power. Though he claims that this “is no representative” his reasoning is that it such an assembly will often be mute where its voice is needed and so it is “unapt…for a government, especially in time of war” (Leviathan, 82–83). I take Hobbes here as saying that such a government is inadvisable due to the cited limitation, not that it is actually impossible. Majority opinion would seem, then, to be not the only available procedure, but the default procedure where no other procedure has been explicitly authorized.


Leviathan, 82.


On the Citizen, 77.


In fact, there are more civil persons than that (see below).


Recall that the assemblies that fall in the scope of the problem of the unity of the representative assembly are those that are free to appoint their own officers: sovereign assemblies and private assemblies (lawful or unlawful). In the case of subordinate political bodies, which are constituted to serve the purposes of the sovereign and act only within the domain specified by the sovereign’s orders, it will not be left to the assembly members to authorize either their procedure or their membership.


For instance, suppose the twelve friends have set out to get some honey for their tea party and that they have all agreed that they will get whichever sort of honey the majority of them prefer. Unfortunately, they are evenly divided with respect to preference for clover or mesquite honey. To resolve the tie, they agree that they will ask Mr. Billingsley his preference. Since Billingsley – though a very fine bee keeper – has notoriously bad taste, they resolve to break the tie in favor of whichever sort of honey Billingsley does not prefer. Mr. Billinsgley prefers mesquite honey; so, the group purchases clover. Here Billingsley’s judgment is authorized as an input, though he is not a member of the group.


This is a special case of the problem about action: the assembly’s speaking to the body it represents is one of the acts it can perform if that problem is solved.


Leviathan, 88.


Leviathan, 100.


The purpose of taking the office view, in the quote from Hobbes above, is to ensure the preservation of the sovereign assembly beyond the natural lives of its members. This might also be a concern in private systems, but it is irrelevant in subordinate political bodies, which are constituted and dissolved at the sovereign’s pleasure. But, again, these are precisely those assemblies that evade the problem of the unity of the representative assembly and so are not of concern for present purposes.


See Martinich, The Two Gods of Leviathan, 357–8 responding to Q. Skinner, “Conquest and Consent: Thomas Hobbes on the Engagement Controversy,” in G.E. Aylmer (ed.), The Interregnum: The Quest for Settlement, 1646–1660 (London: Macmillan, 1972).


For instance, see (Leviathan, 88) in which Hobbes argues that the subjects cannot change their form of government.


For instance, with respect to questions of succession to the membership of the assembly, where the procedural representer has been authorized by the larger multitude, its procedure for succession must either be adhered to or altered by applications of the assembly’s decision procedure. But where the assembly is authorized, the assembly has a much freer hand in determining successors to its offices, as it has fewer constraints on altering its procedural representer.


Here is the text of the sovereign-making covenant: “I authorize and give up my right of governing myself to this man, or to this assembly of men, on the condition, that thou give up thy right to him, and authorize all his actions in like manner” (Leviathan, 88). I take the use of ‘man’ here to be synonymous with ‘person’, but if one takes it as synonymous with ‘human’ the only consequence for my view would be that the procedural representer cannot be authorized as sovereign, only the assembly it unifies can. See below for independent reasons to think that authorization of the assembly is the preferable case, anyhow.


While one might think that the assembly is dissolved by such a move, I think the more fitting analogy here is with representation by conquest: The representer is overthrown by another, to which the assembly binds itself (with majority opinion being implicitly authorized in the interim).


Some odd procedure might not rule in favor of the unanimously held opinion (in much the way a human representer might disagree with all his counselors). But since the procedure must give predictable outputs from a given set of inputs (see Section 3), when the assembly members are in unanimous agreement, they can coordinate to supply the inputs that will yield the desired results.


Here the procedure has a status somewhat akin to an elected monarch. An elected monarch, Hobbes says, is not properly a sovereign, though he does exercise the powers of the sovereignty in practice. Rather the true sovereignty lies with the people that elected him (Leviathan, 100). However, this is a concern only about the sovereignty, and the procedural representer isn’t supposed to have sovereign power on the present view, even if the assembly is.


The transitivity breaks down because the wills/judgments of the assembly members determine the will and will/judgment of the procedural representer (see above, this section).


This worry is particularly pressing for sovereign assemblies because it is a problem that threatens the stability of the civil person represented by the assembly (by undermining its judgments). Where this civil person can dissolve without much negative impact we needn’t worry much about the instability introduced by disputes over the implementation of the procedure. So, in this Section I will speak specifically of sovereign assemblies and procedures. But my remarks will apply to any case of non-sovereign rule by assembly/procedure, where we are concerned with stability. (Such stability will, of course, be bought with a greater curtailing of the scope of private judgment in the relevant domain, along with other freedoms – e.g., the right of exit.).


See Leviathan, 89, 108–110.


See Leviathan, 125, 180–181.


Leviathan, 378.

If the inline PDF is not rendering correctly, you can download the PDF file here.